No abstract

The chicken farm is a typical labor-intensive production environment, and the mechanization of egg picking work is one of the development directions of the chicken industry. This article uses a camera as a sensor for visual detection. Given the limited computing resources of the robot, we improve the feature extraction part of Mask R-CNN network to reduce the memory loss of parameters and speed up the detection process. The experimental results show that compared with the classic method and the Mask R-CNN basic algorithm, the method in this paper has a higher recognition rate that can better support egg picking robots in egg recognition and pose estimation.

No abstract