Real-Time Instance Segmentation and Point Cloud Extraction for Japanese Food

機関リポジトリ 機関リポジトリ (HANDLE) オープンアクセス

説明

Innovation in technology is playing an important role in the development of food industry, as is evidenced by the growing number of food review and food delivery applications. Similarly, it is expected that the process of producing and packaging food itself will become increasingly automated through the use of a robotic system. The shift towards food automation would help ensure quality control of food products and improve production line efficiency, leading to reduced cost and higher profit margin for restaurants and factories. One key enabler for such automated system is the ability to detect and classify food object with great accuracy and speed. In this study, we explore real-time food object segmentation using stereo depth sensing camera mounted on a robotic arm system. Instance segmentation on Japanese food dataset is used to classify food objects at a pixel-level using Cascade Mask R-CNN deep learning model. Additionally, depth information from the sensor is extracted to generate a 3D point cloud of the food object and its surroundings. When combined with the segmented 2D RGB image, a segmented 3D point cloud of the food object can be constructed, which will help facilitate food automation operation such as precision grasping of food object with numerous shapes and sizes.

収録刊行物

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ