スカイロジック List of Products and Services

This is an inquiry from a manufacturer of industrial machinery asking if inspections can be conducted on a conveyor. 【Inspection Settings and Results】 Using DeepSky's inspection capabilities, it was possible to perform evaluations on the conveyor. The left image shows the process of enclosing the area of interest, referred to as "annotation." The right image displays the "detection frame." By separating the settings for overall defects, such as oil adhesion and roughness, from partial defects like scratches, drips, and bumps, we achieved good evaluation results.

• Image Processing Software

We received a request for a free evaluation from an industrial machinery manufacturer to check if there are any omissions in the stationery set packed in the bag. We will inspect whether there are eight types of stationery inside the bag. Although this time it was stationery, this is an example of an application that can be used in various industries, such as checking if assembly parts are complete or if the instruction manual is included. 【Inspection Settings and Results】 Detection was possible as long as the characteristic parts of the work were not hidden. However, there were cases where detection was not possible if the bag was directly under the reflection, so it is stable to place it under the camera with as few wrinkles in the bag as possible. We were able to distinguish the eight types of work that should be inside the bag effectively. The image shows the inspection screen displaying the judgment result table. In the right image, the missing "instruction manual," labeled "B," is highlighted in red.

• Image Processing Software

In the frozen food manufacturing industry, similar to other industries, there is a process of packaging during shipping that includes labeling the product name and date, as well as securing it with PP bands. This time, we conducted an inspection during the shipping process. We assessed four locations based on three criteria: the printing in two places and whether the PP band was correctly secured. 【Inspection Setup and Results】 Using deep learning image processing, we were able to distinguish between OK and NG cartons as shown in the left image. In the left image, the yellow detection box indicates the date printing, the green detection box indicates the product name, and the light blue detection box indicates the detection of the PP band. The inspection points can overlap in this way. To simulate the actual environment, we placed a bar in front of the subject for evaluation. Since it is difficult to recognize the product name if it is obscured from a direct side view, we angled the camera to ensure that more than half of the carton’s text was visible during the shooting and inspection.

• Image Processing Software

Social enterprises and social entrepreneurs have been active in various places recently. We received an inquiry from a manufacturer that operates CSR as a business. This time, the evaluation is about counting goldfish in a tank. [Inspection Settings and Results] The image on the left shows the annotation process, which involves enclosing the objects we want to find. The image on the right shows the detection frame in green counting the goldfish. I cropped the video you sent and tested it with software equipped with AI called DeepSky. As a result of the verification, it seems possible to detect the goldfish. However, detection was not possible when they were hidden behind the hose or when two fish overlapped.

• Image Processing Software

We are considering the introduction of various inspections from a FA control equipment manufacturer. We received an inquiry about automating inspections for "presence of adhesive," "fit claw engagement status," "thermistor adhesive application confirmation," "presence of ring components," "nameplate," "thermistor bending," "presence of screws," "excess/insufficient solder," "presence of lead wires," "solder rise," and "quality of thermal welding." [Inspection Settings and Results] Inspection is likely possible for all inspection items. The initial verification report will be a simple evaluation based solely on pass/fail judgment. The basic concept of image processing using deep learning is to "find what has been taught." For example, when performing "thermistor adhesive application confirmation," you would specify the condition of good products (left) and defective products (right) and conduct learning to distinguish between the presence and absence of adhesive. When considering implementation, please utilize equipment loan services, and ensure that the customer confirms the actual feasibility of inspections.

• Image Processing Software

This is a communication from a trading company we have been dealing with for some time. "Foreign substances" will be detected. We have received inquiries from various customers, including industrial equipment manufacturers, trading companies, and end users. 【Inspection Settings and Results】 When we actually applied image processing, there was some variation in detection depending on the angle, but foreign substances of about 2mm could be detected within a field of view of approximately 250mm. Very small foreign substances, such as tiny black dots or hairs, could be detected within a field of view of about 50mm. 【Software and Equipment Used】 Software Used: DeepSky Learning Field of View: 50mm Minimum Size of Inspection Target: 2mm Number of Inspection Points: 1 location to find foreign substances across the entire screen Camera Resolution: 1.3 million pixels Lens Focal Length: 8mm Lighting: Indoor fluorescent lights

• Image Processing Software

We received sample images of non-woven fabric from various manufacturers of textile products and non-woven fabric. This is a request for a free evaluation to find foreign substances and dirt. Sending actual samples has become stricter recently due to compliance and security concerns. If you can send samples to our company, we can verify and provide guidance on aspects such as lighting, cameras, lenses, and the distance to the work. 【Inspection Settings and Results】 Using the 23 images sent, we were able to detect defective areas. The left image shows the annotation work of enclosing the areas we want to detect with rectangles. The right image shows the detection frames of the inspection results. Out of the 23 images, 16 were used as training images. (There are 21 images, and we processed 2 similar NG images at our company.) We evaluated a total of 23 images, consisting of 16 training images (2 OK / 14 NG) and 7 untrained images (1 OK / 6 NG), and we are reporting the results.

• Image Processing Software

In food manufacturing, the presence of hair can lead to significant incidents. For inquiries regarding food or raw items, we may ask you to send images. In such cases, if you could provide photos with good lighting and focus to clearly capture the area of concern (defect), we can conduct an inspection. This time, we received an image with hair placed on frozen food. 【Inspection Settings and Results】 In the pattern of frozen food, the detection results were favorable. When in a frozen state, the overall appearance is whitish, which provides good contrast with black hair. However, I believe that even at room temperature, there is a possibility of detection if we simply increase the number of training images. By increasing the number of training images and learning steps, DeepSky's strength lies in its ability to improve accuracy through continued learning. 【Software Used】 Software Used: DeepSky Learning Version Number of Inspection Points: 1 location to find hair from the entire screen

• Image Processing Software

The grain dryer manufacturer experienced the web trial of our inspection software DeepSky, which uses AI (deep learning) from our website, but did not obtain satisfactory inspection results. We provided guidance on the method for generating parameters. While personnel familiar with deep learning inspections can conduct inspections smoothly, it is natural that first-time users may not achieve the desired results. If you have any questions about unsatisfactory results or the settings, please feel free to contact us. 【Inspection Settings and Results】 We enclosed the target areas to be detected in frames and labeled them by type. This time, we used seven pieces of training data consisting of five types: the numbers 1 to 4 and one without a number. All images correctly identified the targets. 【Software Used】 Software used: DeepSky Number of inspection points: 2 locations, recognizing the numbers in two areas on the screen.

• Image Processing Software

This is an inquiry from a manufacturer of resin-molded parts for automotive interior components, with whom we have had transactions in the past. Detecting black foreign objects or parts on a black background has been considered a challenging inspection task until now. The image inspection software DeepSky, which uses AI (deep learning), is capable of recognizing such similarly colored parts and foreign objects. 【Inspection Settings and Results】 Using a 1.3-megapixel camera, we captured approximately 1/4 of the overall field of view and conducted learning and detection. As a result, it was possible to identify the types of clips and the number of tackers under the verification environment. However, for items like fasteners attached to sloped surfaces where tackers are difficult to see, detection is likely to be challenging, so adjustments to the camera angle will be necessary. We categorized the labels into 11 types (naming and registering the items we want to find). We conducted simple tests after training with 13 teacher images for 1600 steps. 【Software Used】 Software used: DeepSky

• Image Processing Software

We received an inquiry from a manufacturer of containers and packaging materials regarding the presence or absence of transparent varnish. Detecting the presence of transparent varnish is difficult with conventional image processing methods like EasyInspector, and if we were to make a suggestion, it would be to use a software called DeepSky that utilizes deep learning. 【Inspection Settings and Results】 Since we only have two images for training data, the results will be for reference only. However, by training the model on images with and without varnish, there is a possibility of distinguishing the presence of varnish as shown in the images. The images you provided were likely taken under stable conditions regarding lighting, positioning, and camera distance. We would like to request an increase in the number of NG and OK images using the same imaging method for further validation. 【Software Used】 Software used: DeepSky

• Image Processing Software

A sample product has arrived from the frozen food manufacturer. It has been considered difficult to inspect food items due to their unstable shapes. The AI (deep learning) inspection software DeepSky can demonstrate its strengths against various shapes of parts and defects, and it is easy to set up. Please try the "DeepSky Learning Service." 【Inspection Settings and Results】 We were able to determine the presence or absence of ingredients (eggs) in room temperature fried rice. Although there were some false detections because we could not create an image with the eggs perfectly removed, there was a clear difference in the quantity detected between the presence and absence of eggs. In this verification, we are judging OK/NG based on the premise that there may be some false detections; if the quantity found is small, it is NG, and if it is large, it is OK.

• Image Processing Software

The introduction of appearance inspection systems is rapidly increasing even among food manufacturers. It has been considered difficult to conduct appearance inspections due to unstable shapes, but we expect inquiries to continue to rise due to technological advancements and labor shortages caused by population decline. It is our mission to contribute to the manufacturing industry, which is directly linked to our comfortable lives. 【Inspection Settings and Results】 A total of 31 images were used as training data, consisting of 10 OK and 29 NG samples. As shown in the left image, the areas of interest to be detected were framed and labeled by type. The label names were set as follows. There were no false detections in the judgment based on the training data. Among the 61 non-training images, there were 15 instances where NG was judged as OK. For workpieces with unstable shapes, increasing the number of training images and conducting additional learning can improve detection accuracy. 【Software and Equipment Used】 Software Used: DeepSky Learning Version Number of Inspection Points: 1 point searching for defects across the entire screen Camera Resolution: 1.3 million pixels Lens Focal Length: 12mm

• Image Processing Software

In visual inspections conducted by workers, there is often variability in the criteria for determining "good products" and "defective products" among inspectors, and the more ambiguous the criteria, the more inconsistent the inspection results tend to be. Please consider a stable and efficient inspection using an image inspection system. The image on the left shows the environment with the camera, lens, and lighting. The image on the right depicts the task known as "annotation," which involves framing the areas to be detected. By "training" the framed areas, we were able to detect the target objects within the inspection images. The inspection items were placed on a rotating platform, and detection verification was conducted while rotating. During the verification, it was assumed that the inspection would take about 60 seconds per item. As for dust, it could not be visually confirmed as defective, so settings and detection could not be performed. If defective parts cannot be photographed, the inspection becomes challenging. Creative solutions are needed for capturing images.

• Image Processing Software

In food manufacturing, irregularly shaped work (ingredients) is commonplace, and even minor defects or foreign matter can lead to significant accidents. This time, we have received a request for a simple inspection of food. The ongoing societal issues of a declining workforce and the aging of skilled workers show no signs of stopping, and efforts to support productivity improvement are increasingly in demand. Image inspection technology has gradually become more sophisticated. Please consider the introduction of image inspection to maintain the Japanese quality that allows us to consume safe and secure food. 【Inspection Settings and Results】 This is a verification based on the images provided. The areas of interest to be detected were enclosed in frames and labeled by type. This time, we annotated the images into four categories: "discoloration," "flower buds," "length," and "length 2." Since the verification was conducted by dividing the existing images into labeled and unlabeled data, the accuracy was not very high, but it can be inferred that increasing the sample size will improve accuracy to some extent. 【Software Used】 Software used: DeepSky Learning Edition

• Image Processing Software

This is an inquiry from a food manufacturer regarding whether there are any foreign substances in food products made from beef hide. We tested the provided images to see if we could detect three types of foreign substances: "corn," "straw," and "cow hair," using a simple inspection. Our inspection software has a track record of over 2,000 image inspections. We have numerous inspection cases published, including metals, plastics, food, electronic circuit boards, and pharmaceuticals, so please check our website to see if there are similar cases to the inspections you are considering. 【Inspection Settings and Results】 As a result of the verification conducted with the images you sent, it was possible to detect "corn," "straw," and "cow hair." However, it is necessary to conduct sufficient verification in anticipation of smaller detection items or different patterns (features) that may arise. We determined that detection would be difficult with full-width images due to the small size of the detection targets. This time, we conducted the verification with half-width images. We labeled the annotations as "corn" for corn, "straw" for straw, and "cow hair" for cow hair. We trained the model using 90 annotated images as teacher images. It was set up so that if any one of corn, straw, or cow hair was detected, it would be deemed a failure.

• Image Processing Software

The manufacturer that produces various sample books has also considered our inspection software. In cases like wood grain patterns, there are often discrepancies in color, texture, and design, making it difficult for traditional rule-based image inspection methods. Our website offers a free web trial of the inspection software. You can experience the inspection software that uses AI (deep learning) and make inquiries based on that experience. [Inspection Settings and Results] The images are annotations. By using software that employs AI (Deep Learning) for inspection and training it on the areas to be detected, the software itself adjusts its setting parameters and grows to recognize them. As a result of testing with the samples you provided, there were some misjudgments, but we achieved 100% detection with the trained images. Since the sample products used as teacher images were only three, there were instances of false detection for "similar colors" and "variations in wood grain." By increasing the number of teacher images, it is possible to learn various patterns of wood grain (features) and improve the accuracy rate. Additionally, capturing images separately from the left and right (with two cameras) and enlarging them is also an effective method.

• Image Processing Software

In the food industry, contamination with foreign substances has long been a significant issue. This time, we conducted inspections using a software called DeepSky, which utilizes AI (Deep Learning). The image in the upper left is called an annotation, where we train the software to recognize specific areas (foreign substances) by adjusting its own setting parameters. The software detects "hair," "plastic," and "insects" in the images. However, it was unable to identify the insect mixed in with the sesame seeds on the rice. It is necessary to capture a clear distinction between the sesame and the insect. Software used: DeepSky

• Image Processing Software

We received an inquiry regarding the detection method of DeepSky from a pharmaceutical company we have been in contact with for some time. 【Test Settings and Results】 Image B is one of the four divided parts. The resolution is simply one-fourth. Assuming the area of defects in the entire field of view is 1, the area of defects in the entire field of view of Image B would be 4. Since the "target object" is significantly larger in the overall image, it becomes easier to detect in Image B, even though its resolution is lower. This is why the importance of resolution itself is lower in deep learning. 【Software Used】 Software used: DeepSky

• Image Processing Software

This is a request for evaluation from a manufacturer of audio systems such as speakers. Before the inquiry, line workers were conducting visual inspections of about 10,000 units a day. Due to the black resin and dark adhesive, the work had characteristics that made it very difficult to see, both visually and with cameras. 【Inspection Setup and Results】 Since it was difficult to conduct inspections with EasyInspector, we used a product that utilizes deep learning called DeepSky. Initially, we tried training with the existing images, and detection was possible; however, the number of data points was small, and the images were very difficult to inspect, so I believe more detailed verification is needed in terms of accuracy. Nevertheless, detection was achieved to some extent, and I am reporting the above image as documentation of that process. 【Software Used】 Software used: DeepSky

• Image Processing Software

This is an inquiry regarding the appearance inspection of shumai from a food manufacturer. The food industry has many products with irregular shapes, making it difficult to conduct appearance inspections using traditional rule-based methods. Recently, the number of inspection software using AI (deep learning) has increased, and there are more case studies in this industry. Inspection settings and results For the 13 patterns of samples we received this time, it was possible to distinguish between OK and NG using DeepSky. An overview of the judgment results is provided below. Since there was a concern that the sample products might deteriorate over time, we captured images with a camera of approximately 5 million pixels in advance, and during the actual verification, we downsized those images to about 1.3 million pixels for image processing. The field of view and WD (the distance from the inspection object to the lens tip) for actual implementation will be considered separately. Our inspection software has a track record of over 2,000 image inspections. We have numerous examples of inspections for metals, plastics, food, electronic substrates, pharmaceuticals, and more, so please check our website to see if there are similar cases to the inspections you are considering.

• Image Processing Software

This is a simple evaluation to detect the "black spots" dirt on workpieces, based on inquiries from manufacturers of precision cleaning agents and car chemical products. The priorities are (1) the presence or absence of lot printing, (2) the proper position of lot printing (front and back of the bottle), and (3) the automation of inspection operations for label dirt. Our company generally offers two types of inspection software: the traditional rule-based image inspection "EasyInspector" and the AI-based "DeepSky." This time, we will use DeepSky, which specializes in detecting black spot-like dirt without fixed positioning, assessing the entire screen. 【Inspection Settings and Results】 As a result, we were able to detect the "presence or absence of printing" and "dirt." We registered multiple images (5 images during verification) of the "dirt" area using the inspection software "DeepSky," performed "annotation," and conducted "learning." Since there was only one type of dirt sample, we reproduced dirt in one additional location (by applying transparent tape and recreating black spots on top of it).

• Image Processing Software

Press felt parts for automotive components may vary in the presence or absence of holes and the positioning of parts due to differences in specifications. However, in the case of similar products, accidents have occurred where similar but incorrect items were shipped because they looked almost identical. This time, we will verify whether the installation position of the clips is correct. 【Inspection Settings and Results】 We were able to distinguish between two types of clips in the overall view. It seemed challenging to determine the presence of the fastener's tacker or the text on the label from the overall view, so I suggested that using two cameras for one workpiece would be a more practical approach. The image shows the setup for inspecting whether the correct type of clip is in the correct position. The outer frame indicates the correctness of the installation position, while the inner frame determines whether the part is correct. 【Software Used】 Software Used: DeepSky

• Image Processing Software

We received an inquiry from a customer who manufactures automotive parts and industrial machinery regarding the settings of DeepSky. They asked, "I thought the system was searching the entire screen based on the characteristics of the annotated area I set, but do you also remember any tendencies about where it tends to appear in the field of view?" 【Inspection Settings and Results】 We arrange squares of the same color and size and annotate only the upper part for training (see the image in the top left). In the image on the top right, we can find only the upper part. In the lower left image, we can also find only the upper part. However, in the right image, we could not find anything from the lower part. If the system can only find data that closely matches the teacher due to overfitting, the situation changes a bit. However, the position of the items we want to detect, such as defects, greatly affects inspection accuracy. It is necessary to use teacher images with various positions, orientations, and angles. As of 2021, DeepSky has been equipped with a teacher image augmentation feature. We can augment images of defects that rarely appear by flipping or rotating them in all directions and using brightness or Gaussian noise.

• Image Processing Software

We received multiple verification requests from a truck parts manufacturer. In the free sample evaluation, we first conduct a simple verification using the samples or images provided and report the verification results. If requested, we recommend conducting tests (feasibility verification) assuming actual operation, where we evaluate processing time and judgment accuracy. If feasibility verification is to be conducted, you can choose to continue with our company (for a fee) or use our loaned equipment for verification at your company. [Inspection Settings and Results] We verified whether we could detect burrs using DeepSky with the samples provided. As a result, we detected burrs in 54 out of 55 points, with one point not detected. All other burrs were successfully detected. With DeepSky, we can set the undetected images as additional training images, allowing us to gradually improve accuracy with each occurrence of undetectable defects after operation.

• Image Processing Software

In traditional rule-based image inspection, there are cases that are difficult to handle, but with "DeepSky," which uses AI (deep learning), we can apply it to various inspections that were previously impossible. Many people associate visual inspections with small and medium-sized enterprises, but the same applies to large companies, where surprisingly, automation has not progressed in certain areas. While automation in inspection lags behind production, issues such as the aging of inspectors and labor shortages have become problematic. We hope to contribute to efficient production through the automation of inspections. [Inspection Settings and Results] In the verification of lemon grade classification using the provided images, we were able to distinguish four grades with approximately 86% accuracy. Please consider this value as a reference due to the small sample size. We believe that increasing the sample size will lead to improved accuracy. With the version upgrade in 2021, we were able to incorporate a data augmentation feature for training images. This allows for efficient and stable inspections even with a small number of samples. We will continue to add various convenient features and develop user-friendly inspection software from a practical perspective.

• Image Processing Software

We received an inquiry from a manufacturer of various machine parts regarding bearings. They currently monitor the treatment of waste liquid through visual inspection, and it is said that when aggregation occurs from the state shown in the left four images below to the state in the right four images, it is considered acceptable. Although there is a roof, it is not an indoor environment, and the imaging conditions for inspection change depending on the time and weather. 【Inspection Settings and Results】 We set 24 images, including the ones above, as training images, annotated them all, and were able to make judgments. With the version upgrade in 2021, we were able to incorporate a data augmentation feature for training images. We will continue to develop inspection software that is user-friendly and practical by adding various convenient features. The strength of AI (deep learning) image inspection software compared to traditional rule-based image inspection software is its ability to perform inspections even with differences in brightness and imaging environments. Inspection software can be beneficial in various industries and processes. 【Inspection Software】 Software used: DeepSky

• Image Processing Software

The photos you provided made it difficult to conduct inspections due to the shooting environment. While I was able to make a judgment as a simple inspection, it is particularly necessary to carefully consider the lighting, camera, and lens setup when dealing with glossy workpieces. [Inspection Settings and Results] You sent a total of 9 images, of which 6 were used as training images. Each of the 6 images was divided into 60 segments using image segmentation software, and 23 segments with scratches were used for training. The remaining 3 images were used for testing, and these were also divided into 60 segments before inspection. I also tried to detect dents, but I was unable to do so. It is thought that the difficulty in visually identifying certain areas may be the reason for the unsuccessful training, and I believe that defective detection can be achieved depending on the imaging environment. [Software Used] Software used: DeepSky

• Image Processing Software

This is a case that we continued to consider after a simple verification. In image inspection, the shooting environment is very important. If we can accurately capture defects, inspection is possible with a high probability. It is necessary to prepare the inspection environment within the operational range of the camera, lens, and lighting. [Inspection Settings and Results] While continuing our discussions, we found that using a backlight allowed us to clearly capture the entrapment (defect), which is expected to improve the detection rate, so we were able to provide a good report. The device used this time is a tracing table that costs several thousand yen, so I believe that sufficient effects can be obtained even with lighting that is not specifically for inspection, as long as it is a surface light source. Our company website offers a free web trial of the inspection software. You can actually experience the inspection software that uses AI (deep learning). [Software Used] Software Used: DeepSky

• Image Processing Software

A manufacturer of industrial equipment, with whom we have had a relationship for some time, has been actively inquiring about "DeepSky" since its release. Our company is accepting verification and support requests from our technical staff on a daily basis. If you have any issues, questions, or uncertainties during operation, please feel free to contact us at any time. 【Inspection Settings and Results】 Out of 42 images, 21 were used as training images. We inspected a total of 42 images, consisting of 21 training images and 21 untrained images. The result was that we were unable to detect only one untrained image, while all others were successfully detected. Images that were blurry or had faint prints were generally able to be judged. If we can stabilize the captured images further, we believe the accuracy could improve. 【Software Used】 Software used: DeepSky

• Image Processing Software

We have received a simple free evaluation to determine the types of tires, such as automotive parts. This time, we tried sending photos, but sending actual samples would provide the most reliable results. In that case, you will need to send several good products and the defective work you want to detect. Please contact us for details. 【Inspection Settings and Results】 The left image shows the annotation process (the task of enclosing the parts you want to teach). The right image displays the OK screen. The tire text is being evaluated by DeepSky. We also have the capability to record images and CSV files, making integration with surrounding systems easy. The number of detectable types that can be registered per product type ranges from 1 to 1000 (here, "types" refers to classification names used during annotation, such as "screw," "desiccant," "tomato," etc. If the number of detectable types exceeds 100, the detection rate may decrease.) 【Software Used】 Software used: DeepSky

• Image Processing Software

At the request of a manufacturer of industrial equipment, we conducted a simple evaluation to determine the harvest timing of strawberries. We used 58 sample images for the inspection. (34 strawberries were classified as OK and 24 as NG) 【Inspection Settings and Results】 As a result, all "harvestable strawberries" were correctly detected. Teacher Images: Correct Judgment 100% (20/20) Incorrect Judgment 0% (0/20) Unlearned Images: 94% (32/38) 6% (2/38) Total: 96% (56/58) 4% (2/58) However, among the strawberries classified as NG, two strawberries that appeared close to OK when viewed by the human eye were mistakenly classified as harvestable. (1) Still pinkish in color, therefore not harvestable... NG1 (Pink) (2) Still white or green, therefore not harvestable... NG2 (White or Green) (3) Strawberries that can be harvested... Harvest OK By creating three types of labels and training the software, it will adjust its own setting parameters and improve recognition. The images are annotated.

• Image Processing Software

We received a direct request from a customer who manufactures medical capsules and were conducting a preliminary verification. However, since our company does not sell FA machines, we decided to proceed with the verification through the manufacturer of the control equipment that the customer is dealing with. Our company is located in Hamamatsu City, Shizuoka Prefecture, and we have many industrial machine manufacturers as clients, primarily in the automotive industry. We have developed easy-to-integrate inspection software that is sold outright, and we have received numerous repeat orders from manufacturers of industrial machines and systems. 【Inspection Settings and Results】 As a result of processing the images received, the first image was detected normally. The second image had one false detection and one non-detection. After retraining, it was able to detect normally. After verification, we will conduct a demo at our company to explain the inspection environment and operation methods, and we will lend out demo units so that you can actually check the setup methods and inspection accuracy. 【Software Used】 Software used: DeepSky

• Image Processing Software

We conducted a simple inspection of the extent of food overflow in trays from a food manufacturer using the rule-based "EasyInspector" and reported our findings. However, detecting overflow on trays with a blue background and white patterns proved difficult, so we decided to re-verify using "DeepSky," which utilizes AI (deep learning). Our company website offers a free web trial of the inspection software. You can actually experience the inspection software that uses AI (deep learning). 【Inspection Settings and Results】 We created several patterns of defective conditions using the samples we have on hand and were able to detect ingredient overflow by training on those images. However, in this case, various patterns of ingredients and defects are expected, so a large amount of training data is necessary to achieve sufficient detection results. With the version upgrade in 2021, we were able to incorporate a data augmentation feature for training images. We will continue to add various convenient features and develop user-friendly inspection software from a practical perspective. 【Software Used】 Software Used: DeepSky

• Image Processing Software

We received a contact from a manufacturer considering a poka-yoke device for forgetting to paint rubber hoses. When there is identification paint (white, yellow, red, green) on the end face of a black rubber hose approximately Φ20, it will indicate OK; if there is no identification paint, it will indicate NG during the inspection of body parts. Our website offers a free trial of inspection software online. You can actually experience the inspection software that uses AI (deep learning). 【Inspection Settings and Results】 It is possible to determine that there is one blue mark on the inspection screen. In the case of red, the count for "red" will be one, and if there is one each of white and green, they will be counted separately. Since we have trained the system to recognize the absence of marks as "none," the above image is judged to have one "none." At this stage, we are simply counting how many of each there are, so if there is more than one "none," we will set the system to determine it as a failure. 【Software Used】 Software used: DeepSky

• Image Processing Software

This is a free simple verification based on the sample images you provided. It is from a manufacturer of industrial control equipment with whom we have had transactions in the past. In image inspection, the imaging environment is a very important point. This time, the accuracy of the inspection varied depending on the captured images. 【Inspection Settings and Results】 We were unable to obtain good results with the images as they were, so we cropped only the areas to be evaluated and conducted the verification. As a result, in the case of the images we verified this time, if we could set the field of view appropriately for the judgment of the paper tube, it was possible to make a judgment. Since there were few OK images, we created mirrored images to use for verification. We used a total of 20 images for training, consisting of 10 OK images and 10 NG images. For the validation of the judgment, we used 2 OK images and 2 NG images that were not used in the training, for a total of 4 images. 【Software Used】 Software used: DeepSky

• Image Processing Software

We will measure the "diameter" of the particles of high-silicon material based on the inquiry from the trading company. This will be verified using the images provided, and there is a request to inspect using images from an electron microscope. Our company accepts verification and support from technical staff on a daily basis. If you have any issues, questions, or uncertainties during operation, please feel free to contact us at any time. 【Inspection Settings and Results】 By using the "Dimension and Angle Inspection" function of EasyInspector, we were able to determine the diameter at one location in 0.11 seconds. Since this was a simple report on whether it could be done, actual operation will require positioning of the inspection items. We offer a "60-day trial service for inspection software" that can be downloaded from our website. Additionally, EasyInspector can perform inspections using JPG/PNG/BMP images. 【Software Used】 Software Used: EasyInspector310 Number of Inspection Points: 1 The current 'EasyInspector2' MS (MeaSure) package can perform inspections for [Position and Width Measurement] and [Angle Measurement].

• Image Processing Software

We received sample images from an automotive parts manufacturer. We will report whether the dimensions are within tolerance through a free simple verification. In the inspection process, automation is being promoted to pursue stable inspection accuracy and cost reduction. 【Inspection Settings and Results】 By using EasyInspector's "Dimension Angle Inspection" function, we were able to confirm the tolerance of a dimension at one location (left settings screen). The inspection cycle time is 0.62 seconds. We are measuring the edges of the hole and flat surface in inspection frames 001 and 002 (right detection screen). The software can calculate the difference in the X coordinate values from 002 to 001, and it is possible to convert the dimensions to mm. In the free sample evaluation, we will first conduct a simple verification using the received samples or images and report the verification results. In the simple verification, we will evaluate whether the desired detection/judgment is possible with our in-house equipment. If detection or judgment is possible in the simple verification, we recommend conducting a test (feasibility verification) assuming actual operation, and evaluating processing time and judgment accuracy. If conducting feasibility verification, you can choose whether to continue with our company (for a fee) or to use our rental equipment for verification at your company.

• Image Processing Software

This time, we received an inquiry from a contract analysis company. Work size: approximately 100mm × 100mm. Seal size: approximately 10mm (width). We created a model and conducted a simple verification. In cases where we receive inquiries, there are many instances where we cannot send sample products or sample images due to compliance and security concerns. [Inspection Settings and Results] Using EasyInpector's "Dimension Angle Inspection," we referred to the materials you provided and measured the tolerances with the seals applied by our company. This time, we measured only one location for each seal, but it is actually possible to measure four points simultaneously. We used a 5-megapixel camera (2592 × 1944) with a field of view of 180mm horizontally, capturing images of a 100mm × 100mm metal plate. We measured by applying two seals, each 10mm wide. There was some slight shaking during imaging, resulting in an error of 3 to 4 pixels, but detection with an accuracy of 0.5 to 0.7mm in tolerance is theoretically possible. Increasing the camera resolution will improve accuracy further.

• Image Processing Software

This is an inquiry from a metal processing manufacturer that has already implemented our solutions. They are engaged in various business developments and will conduct a simple verification of LED lighting inspection this time. In the free sample evaluation, we will first perform a simple verification using the samples or images provided and report the verification results. During the simple verification, we will evaluate whether the desired detection/judgment can be achieved using our internal equipment. If detection or judgment is possible in the simple verification, we recommend conducting a test (feasibility verification) assuming actual operation, where we will evaluate processing time and judgment accuracy. If feasibility verification is to be conducted, you can choose to continue with us (for a fee) or use our loaned equipment for verification at your company. 【Inspection Settings and Results】 By using the "Presence of Specified Color Inspection" function of EasyInspector, it was possible to determine the pass/fail of 48 LED locations. This time, capturing the light edges clearly was challenging, so we either reduced the exposure or narrowed the aperture to capture the particles as distinctly as possible. After that, we can only search for values that can be successfully detected while tightening the color judgment tolerance. We will summarize and report the setting values for reference, but please note that the environments differ between us and you, so having the same values does not guarantee the same results.

• Image Processing Software

The sports equipment manufacturer has been using our inspection software overseas for some time. This time, we have questions regarding the settings for "matching numbers on label printing," "smudge detection," and "checking for the presence of instruction manuals." We have agents in Malaysia and China, and we also sell inspection software for use in overseas offices. Our inspection software includes basic functions for saving inspection results in CSV and image formats, which are useful for product management across various industries. [Inspection Settings and Results] By using the "comparison with master image" function of EasyInspector, it was possible to inspect the codes and prints on 16 labels in 0.28 seconds. In image inspection, the imaging environment is the most important factor. Regarding the reasons for a dark screen during inspection, please check the following: whether there are fluorescent lights near the inspection target (this is the most common external influence); whether there are moving machines (if there are motors using electromagnetic forces, such as AC servo motors, noise may occur); and whether there is a 100V switch (which can generate electromagnetic noise).

• Image Processing Software

Samples were sent from the comprehensive processing manufacturer. A simple verification was requested to determine whether the brightness at "DC14.4V is OK" and "DC9.0V is NG" can be distinguished. It was requested that if even one LED is dim during the inspection of multiple units at once, it should be judged as NG. 【Inspection Settings and Results】 We conducted an inspection to verify the brightness difference by powering at DC14.4V and DC9.0V, and we are reporting the results. The inspection was possible using the "Presence/Absence Inspection of Specified Color" feature of EasyInspector. We adjusted the lens aperture and exposure time for imaging and conducted the inspection in an environment where differences were easy to discern. By making the entire circuit board the field of view and setting inspection frames for each LED (30 units), simultaneous inspection was possible, allowing us to output which LED was NG. This time, we set the entire circuit board as the field of view and verified by applying power to one LED. The left image shows the pink inspection frame. The right image displays the green pixels detecting the specified color, while the blue color indicates the inspection frame that passed.

• Image Processing Software

Our company is located in Hamamatsu City, Shizuoka Prefecture, which is known for its thriving automotive parts manufacturing industry. Our main customers are manufacturers of various industrial machinery and control equipment, as well as users. We have a track record of providing support for over 1,000 cases and are developing inspection software that is easy to integrate and sold as a one-time purchase. 【Inspection Settings and Results】 By using the "Specified Color Presence Inspection Function" of EasyInspector, we conducted an inspection at one location. We used a circular frame (optional), but by enclosing the entire workpiece and masking the black background and seam areas, we believe it can also be inspected with a standard frame. In the left image, the masking function allows us to set "non-detection pixels" for areas that should not be inspected, which helps shorten the cycle time and eliminate false detections; it is a highly convenient feature. The right image shows the detection results. 【Software Used】 Software Used: EasyInspector710 The current 'EasyInspector2' color package can perform inspections for "Specified Color Presence."

• Image Processing Software

We received an inquiry from a battery manufacturer stating, "There is a white + mark on the cover, and we want to determine if there is any mistake in orientation based on its position." They also expressed a desire to take photos while holding a tablet and camera by hand. 【Inspection Settings and Results】 By using the "Presence of Specified Color Inspection" feature of EasyInspector, we were able to determine the orientation of the parts. We reported on a simple verification. Inspection frames were set up in multiple locations (this time at six locations: including two NG locations). We set "white" as the specified color and conducted the inspection based on the extent to which the specified color was detected within the inspection frame. When using EasyInspector, it is essential to fix the camera. While it is possible to take photos with a tablet, it needs to be secured. 【Software Used】 Software used: EasyInspector710 The current 'EasyInspector2' color package can perform inspections for the "Presence of Specified Color."

• Image Processing Software

The fact that 'there are inevitably oversights by operators during visual inspections' can occur not only in factory production processes but also in other areas. This inquiry comes from an industrial machinery manufacturer, but the verification of check sheets filled out by humans is a common issue across all industries. [Inspection Settings and Results] We printed the data from the check sheet you sent and conducted tests. We verified the presence or absence of checks on paper. By using EasyInspector's 'OCR Pro' feature, we were able to read five holes or symbols in just 0.62 seconds. After a simple verification, we will conduct an in-house demonstration. Additionally, we offer a web video demonstration on our website, a 60-day trial software download, and a free evaluation service for inspection items, all aimed at providing information and support to help customers easily implement the inspection system. We look forward to your inquiries and hope you take advantage of the above services. If you have any requests or questions, please feel free to contact us anytime.

• Image Processing Software