スカイロジック List of Products and Services

With the development and spread of AI, it has become possible to automatically detect foreign objects (such as insects, frogs, anisakis, hair, etc.) in food that were previously considered difficult to identify. AI can discover foreign objects that are often overlooked during quick tasks. ■ Hair contamination in food ■ Confirmation of ingredients in cup noodles ■ Identification of bread types on trays ■ Differentiation of meat cuts ■ Counting packaged sweets ■ Inspection of egg cracks ▼ Preventing oversights with AI to ensure food safety Issues such as damage or rot in agricultural products, ingredient shortages in food, and foreign object contamination cannot be completely prevented even with visual checks, leading to oversights. AI continuously monitors agricultural products and food without fatigue, efficiently reducing the chances of missing anything. ▼ Standardization and labor-saving in sorting Agricultural products have various criteria for checks, such as damage and insect bites, and also need to be sorted based on color and shape. When done by human hands, there is inevitably some oversight and subjective judgment, making it difficult to maintain consistent quality. The AI's vision checks and sorts based on constant standards, contributing to the standardization and labor-saving of sorting processes.

▼Obtaining Anisakis → Until Filming This time, we conducted a verification at our company to find "Anisakis," which we received several inquiries about from customers, using AI. We started by trying to obtain live Anisakis, but it turned out to be more difficult than we expected... We visited several supermarkets and fish shops in the city and made phone calls to request their help. Despite it being an extremely busy time of year, I made a strange request to unknown housewives, saying, "Can you give me live Anisakis for an experiment to find Anisakis with AI (to summarize)?" I am very grateful to the supermarkets that cooperated with us, and to the staff in the fresh fish section who helped us. Thank you very much. We assembled the equipment assuming an operation where we conduct AI Anisakis inspections on a slowly moving conveyor, and if detected, a buzzer would sound to stop the conveyor. Anisakis glows in response to specific wavelengths of ultraviolet light, so we use UV LED light sources and filters that pass specific wavelengths. In image processing, it is also important to capture images that make it easier to find Anisakis through optical manipulation.

I believe there are many instances where people visually check for insects or hair on food flowing on a conveyor. Here, I would like to introduce a system that incorporates AI vision to more reliably detect foreign objects. By controlling the ON/OFF of a 100V power supply with a relay, the conveyor itself is turned OFF when a foreign object is detected, stopping the conveyor and simultaneously sounding a buzzer to alert about the foreign object detection. This system uses DeepSky and Intelligent I/O, with a relay connected to the NG output of the Intelligent I/O. For more details, please check the related links below. 【Tools and Equipment】 ■ PC: Mouse Computer/G-Tune, Core i9, 16GB RAM, RTX 2070 SUPER ■ Image Processing Software: Skylogic/DeepSky DS100K ■ IO Unit: Skylogic/EI-ITIO-T01 ■ 1.3 Megapixel Camera: Daheng/MER-133-54u3c ■ 8mm Lens: M0814-MP2 ■ Camera Stand (Aluminum Frame) *For more details, please refer to the PDF document or feel free to contact us.

We would like to introduce a case where the overflow and presence or absence of bubbles in a tank storing a chicken egg solution were detected in advance. By using the "Presence or Absence Inspection of Specified Color" feature of EasyInspector, it was possible to conduct an inspection at one location. When the liquid is not foaming, a pass indication is displayed with a blue frame as shown in the left image. In the right image, the detected pixels indicating foaming are shown in light blue, and the red frame indicates a failure. 【Software Used】 ■EasyInspector (formerly EasyInspector) The current 'EasyInspector2' color package can perform inspections with the "Presence or Absence of Specified Color." *For more details, please refer to the related links or feel free to contact us.

We received an inquiry from a food manufacturer. In the process where the operator adds ingredients one by one, we check through image inspection whether the oden ingredients are properly placed in the container. Since daikon is added first, there is no risk of forgetting it, so we will verify the other four types: chikuwa, konnyaku, tsukune, and egg. We trained DeepSky on the four types of ingredients: chikuwa, konnyaku, tsukune, and egg, and set it up so that it is considered OK if it detects one of each. There are times when detection fails if the target ingredients are hidden, but it was generally able to detect them when more than half of the ingredients were visible. 【Software and Equipment Used】 Software Used: DeepSky Field of View: Approximately 163 x 130mm Minimum Size of Inspection Target: 40mm Number of Inspection Points: 4 Camera Resolution: 1.3 Megapixels Lens Focal Length: 12mm Distance Between Lens and Product: 330mm Lighting: Indoor lighting

If the side dishes lined up in the supermarket overflow from their trays, there will be concerns about leakage and poor presentation, which may lead to unsold items. This time, we received a request for a simple verification regarding the overflow of food placed in trays from a food manufacturer. By using EasyInspector's "Presence of Designated Color" feature, we were able to inspect one area (the entire screen) in 0.39 seconds. The left image shows the designated color settings, where we set the color of the food that may overflow. The right image shows the mask settings (specifying non-detection pixels), where the specified area will no longer detect the designated color. 【Software and Equipment Used】 Software Used: EasyInspector (formerly EasyInspector) Field of View: Approximately 10 x 8 mm Minimum Size of Inspection Target: 5 mm Number of Inspection Points: 1 Camera Resolution: 1.3 Megapixels Lens Focal Length: 6 mm Distance from Lens to Product: Approximately 165 mm Lighting: Ring Lighting Distance from Lighting to Inspection Item: Approximately 120 mm The current 'EasyInspector2' color package can be used for inspection with the "Presence of Designated Color" feature.

The confectionery manufacturer in this inquiry is packaged in cardboard boxes, just like other industries. We will conduct a simple verification to check whether the manufacturing date is printed on the cardboard box. The EasyInspector software (traditional rule-based image inspection software) that you inquired about is primarily used by customers in the automotive parts and electronic circuit board sectors, making up half of our clientele. Additionally, it is utilized by a wide range of industries for tasks such as assembly verification and inspection of minor product dirt. We plan to use the EasyInspector feature called "Color Presence Inspection." This function allows us to specify an ink color and determine whether it falls within the specified range for a pass/fail judgment. In the left image, the specified color is detected as red, and a passing blue frame is displayed. In the right image, the specified color could not be detected, resulting in a failing red frame. It is also possible to conduct simultaneous inspections at two locations by using two cameras.

We reported the presence or absence of tears in shrink-wrapped products from food manufacturers through a free simple verification process. We would like to continue verifying issues such as deformation, stacking collapse, and flap adhesion, and propose various defect detection methods. In the simple verification, we were able to accurately identify approximately 80% of tears that are easily noticeable by the human eye, and about 50% of tears that are less noticeable, such as those on white products. This time, we conducted the verification with a limited number of samples, and I believe the accuracy was affected by the small amount of training data. The detection accuracy will improve with more data for training. [Software Used] Deepsky

We have received several inquiries from various food production and processing sites. We are now able to handle cases that were difficult for EasyInspector (formerly EasyInspector), such as detecting hair and foreign objects mixed in bento boxes, which further demonstrates the wide range of capabilities of our AI (deep learning) functions. DeepSky allows us to teach it with a broader scope regarding what we want to detect, making it possible to identify foreign objects that do not conform to a specific shape. If there are items that are currently being checked by human eyes and you wonder, "Can this be inspected?" please feel free to contact us.

Recently, we have been receiving inquiries asking if the inspection of agricultural products we are currently conducting can be made a bit easier. Using vegetables provided by local farmers, we tested whether we could: ■ Classify potatoes by grade ■ Detect damage on tomatoes using SkyLogic. ▼ Potato Classification We photographed potatoes classified from A to C from different angles (a total of 66 images were taken), and after creating the training data and conducting inspections, we were able to classify them into grades A, B, and C. (All 30 images were classified correctly, achieving a 100 percent accuracy rate.) ▼ Detecting Damage on Tomatoes Since it was difficult to find damaged tomatoes, we tested whether we could detect damage on one type. Similar to the potatoes, we photographed them from different angles (12 images), created training data, and conducted inspections, resulting in the detection of damage as shown in the images. (All 30 images were inspected, and all damage was detected, achieving a 100 percent accuracy rate.)

Until now, Skylogic has primarily handled image inspection projects focused on industrial products. This is because the nature of industrial products, which involves "mass-producing the same item," matched well with conventional (rule-based) image processing methods. Conversely, it has been difficult to handle items with unstable colors and shapes, even among industrial products, using traditional methods. However, with the emergence of image processing methods utilizing AI (deep learning), items with variability can now be included in the scope of image processing. This is because AI excels at "detecting only what we want to find while ignoring trivial changes." Now, connecting this to the title, when we talk about "items with unstable colors and shapes," we are referring to food products. Issues such as material defects, foreign object contamination, quantity defects, and missing components... even when the subject is food, products produced in factories are likely to face similar challenges as industrial products. However, the difficulty of processing with traditional methods has led to these issues being abandoned. (And I have also given up on them.) Why not try AI?

Do you have any of these concerns? ■ It takes a lot of time for manual work ■ Grade classification is difficult, leading to oversights in visual inspections ■ Relying on hired labor increases labor costs ■ While skilled workers are aging, there are no successors Automate with image inspection software! 1) The defect detection tasks that were done manually can be automated ⇒ Detect scratches, discoloration, dirt, etc., and reduce work time! 2) Difficult grade classification can be easily done by anyone ⇒ Various rules can be used for differentiation! Free yourself from manual work! 3) Operate from a single computer! Start at a low cost ⇒ Significantly reduce implementation costs! Keep the number of workers to a minimum and reduce labor costs! Prevent oversights in visual inspections and enhance quality value! You can start small without the need for large machinery! 【Inspection Examples】 ■ Classification of tomatoes by size ■ Classification of flower types ■ Identification of corn seeds ■ Differentiation of tea leaf types ■ Inspection of orange blemishes ■ Detection of defects in coffee beans *For more details, please see the related links.

Insects, plastic, frogs, anisakis, hair, and other foreign objects are monitored at a speed of about 10 times per second. AI eyes efficiently detect foreign objects on behalf of human eyes and notify with lights or buzzers. When connected to a PLC, it is also possible to stop or discharge the conveyor upon detection of foreign objects. ▼ Easy installation on conveyor belts Even in places where it's difficult to place a PC, the aiforce-mini can be installed and used on the side of the conveyor or control panel. ▼ Detectable foreign objects Insects, plastic pieces, frogs, anisakis, hair, and other items that can be visually confirmed. ▼ Beyond foreign objects Through learning, it can also check for scratches, damage, burns, and chips on the conveyor. 【Product Composition】 - aiforce-mini main unit (AI software pre-installed) - Industrial camera - Lens (specifications are selected based on the inspection target and shooting distance) - USB cable for the camera (3m) * A mouse, keyboard, and monitor needed for setup are optional.

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

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.

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.

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.

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

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.

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

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

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.

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

Seafood and fruits have a rough similarity in shape, but in image inspection, their shapes are not stable, and variations in size and ripeness can change their color and bulge, making them difficult to assess. Our inspection software, DeepSky, excels in inspecting items with slightly different shapes. This time, we evaluated the defect detection of shucked oysters based on an inquiry from a manufacturer that produces automation robots used in food factories. Since it is not the shipping season for oysters, the inspection was conducted using the images provided. 【Inspection Settings and Results】 Upon verifying the sample images you sent, we were able to detect eggs and black spots. The inspection was performed using software that employs AI (Deep Learning). By training the software on the areas we wanted to detect, it adjusts its own setting parameters and learns to recognize them. Our software does not only detect based on color; it also incorporates texture (surface quality) into its judgment. Additionally, there is no need for fixed positioning, allowing for detection even if multiple oysters flow on a conveyor in various positions and orientations.

In the food industry, foreign object contamination can lead to significant accidents. As of 2017, verification was conducted using the conventional inspection software EasyInspector, but now such cases can be easily configured and inspected with DeepSky. Please refer to the article "Stopping the Conveyor After Discovering Insects on Cabbage." 【Inspection Settings and Results】 By using EasyInspector's "Presence of Specified Color Inspection" feature, it was possible to detect different colors from one location (the entire field of view). However, since the inspection was based on submitted images, there were instances where the background gray was similar to the different color, leading to noticeable false detections. While it is possible to set a strict "color judgment tolerance range," it has been suggested that adjustments to background color and lighting are necessary. 【Software and Equipment Used】 Software Used: EasyInspector710 Number of Inspection Points: 1 (entirety) Current 'EasyInspector2' color package allows inspection for the "Presence of Specified Color."

Currently, cases that are verified with DeepSky were previously verified with EasyInspector before the release of DeepSky. This verification involves inspecting the shape of the base of green onions. It includes sorting those with a bulbous, rounded base and many roots. The customer has specific requirements for the camera and resolution. 【Inspection Settings and Results】 We are inspecting the shape of the base of green onions using EasyInspector's "Dimension and Angle Inspection" function and "Damage Inspection" function. We propose various measurement methods across 14 different items. The inspection content is something that our software DeepSky, released in 2020, excels at. Our company aims to solve customer issues as cost-effectively as possible by utilizing the equipment they currently have, such as cameras.