Development of Automatic Recognition and Gripping System of Beekeeping Auxiliary Device Using Stereo Camera and Object Detection Algorithm

초록

Apiculture has been thriving and improving as a part of the agricultural industry. Many advanced tools have been invented to support beekeeping management processes. However, hive handling has always been the most challenging problem for apiarists in daily beekeeping processes. This paper aims to solve the daily beekeeping bee box transport problem by developing an automatic recognition and gripping device based on a stereo camera ranging method and object detection algorithm. Firstly, a hive recognition model was established based on the YOLOv5s detection model. The three-dimensional location of the hive was able to be determined in real-time inference by using the oak-D Lite Stereo Camera. The position information required for the grasping action was provided to the uArm manipulator through simple relation transformation calculation. In this way, the uArm was controlled to operate self-recognition and automatic clamping of the bee box. The cost-effective stereo camera module and lightweight object detection model used in this paper will significantly promote the automation and practical application of apiculture auxiliary devices. © 2022 ASABE. All Rights Reserved.

키워드