Recognition of Specific Parts of Plastic Bottles Using Improved DeepLab v3+
Daiki Ideta, Tohru Kamiya · 2023
In this paper, we focus on the manpower shortage in factories and conduct an experiment to try to automate the process. Factory automation has many advantages, such as reducing labor costs and improving production efficiency. Among them, we focused on the sorting of plastic bottles at a waste disposal plant. Currently, garbage is sent to the landfill without being sorted, and it is sorted by hand. Unlike cans, plastic bottles cannot be easily automated because they cannot be separated using magnets. Therefore, we propose an image processing technique to recognize where plastic bottles are located, and to use a robot arm to sort them, thereby realizing automation. In this paper, we focus on image processing technique based on deep learning. As basic research, we conducted an experiment to see how well the robot can identify a single plastic bottle in an image. We attempted to use semantic segmentation methods for detection, using DeepLabv3+ as the basic model, and improved it. scSE Block and fine-tuning were introduced to improve the accuracy significantly compared to previous studies.