AI image processing selects the color of glass bottles without using special lighting such as UV, recognizing empty cans and PET bottles. It will be a very inexpensive high-performance device!
Color sorting of glass bottles only using AI image processing. Other items such as empty cans and PET bottles are not sorted by color, only glass bottles are selected. By using HALCON, the deep learning training data consists only of the glass bottles we want to pick, and there is no need to train on prohibited items like PET bottles, empty cans, or cosmetic bottles! ⇒ Patent application planned. The image recognition accuracy for prohibited items is nearly 100%. Lighting does not require expensive options like commercially available white LEDs or UV lights. The learning method has been specialized, and the amount of image data used for training has decreased. Colors are sorted into brown, clear, green, etc. Only glass bottles of the specified color are picked by the robot using conveyor tracking. The robot can be flexibly modified to suit the end users, including: Collaborative robots, Parallel link robots, Articulated robots, and many other types. This sorting machine is a culmination of our company's technology, which knows everything about glass bottles! Plus, it uses the world-class image processing HALCON. Completion is just around the corner. Please look forward to it. *HALCON is a pioneering tool that allows for the construction of efficient image processing systems by freely combining over 2,000 high-functionality functions.
Inquire About This Product
basic information
AI image processing uses HALCON's deep learning. Color camera: iRAYPLE or Basler manufactured area camera. Conveyor width: 500 to 1000 mm. Image discrimination accuracy: 90 to 99%. Lighting is only white LED; no special lighting such as UV or near-infrared is used, resulting in a very inexpensive and high-performance glass bottle color sorting machine developed!
Price information
We are developing with a budget set at 15 to 20 million yen per unit. The price will vary depending on the number of robots. We recommend a basic setup of two robots.
Price range
P7
Delivery Time
Applications/Examples of results
Currently under development, aiming for completion within this fiscal year. AI image processing is nearly complete, and collaboration with the robot is in development.
Company information
In the Vision & Robot System Development Department, we develop vision systems for robotic systems and robotic systems themselves. We are involved in the development of optical sorting machines for recycling companies, automotive parts manufacturers, and food manufacturers, as well as the development of appearance inspection systems for automotive parts manufacturers. Our areas of expertise include optical automatic sorting machines, high-speed image processing, AI (deep learning) image processing, appearance inspection of metal parts, and PLC control design. We also develop and design automatic sorting inspection equipment using robots from scratch. For customers considering factory equipment automation, we provide automatic sorting machines, control panel design for equipment, robot implementation, and vision system implementation. 1. We mainly use KEYENCE, HALCON, and HALCON DL for image processing. We also support other image processing solutions, including MERLIC. 2. For PLCs, we primarily use KEYENCE and Mitsubishi Electric products (we also handle other manufacturers' PLCs). 3. We use robots from DENSO WAVE, Mitsubishi Electric, FANUC, and Yamaha Motor. 4. For cameras, we use color/monochrome area cameras and line scan cameras, including Basler cameras, JAI UV cameras, JAI prism spectroscopic cameras, and KEYENCE cameras.
