基于深度学习的肉鸽精准饲喂机器人的研制与试验

Development and Experiment of Precision Feeding Robot for Pigeons Based on Deep Learning

针对肉鸽饲喂过程自动化程度低、劳动力成本高且专用饲喂智能技术与装备匮乏等问题,根据笼养肉鸽的生长习性,基于深度学习视觉技术,本文研制了一种适用于笼养环境的肉鸽精准饲喂机器人。研制的机器人包含底盘导航模块、运算控制模块、机械臂投喂模块、视觉感知模块,具有作业环境视觉三维感知定位、自主导航、自主识别鸽笼中的笼养信息并进行精准饲喂等功能,实现了在鸽舍环境中自主饲喂作业,平均饲喂成功率为92%。肉鸽精准饲喂机器人有望逐步代替人工饲喂作业,能有效降低行业劳动力需求,提升肉鸽养殖产业的效能。

Aiming at the problems of low degree of automation in the feeding process of pigeons,high labor cost and lack of special feeding intelligent technology and equipment,according to the growth habit of caged pigeons and based on deep learning visual technology,this paper developed a new method suitable for caged environment. pigeon precision feeding robot. The developed robot includes a chassis naviga-tion module,an arithmetic control module,a robotic arm feeding module,and a visual perception module. It has the functions of visual three-dimensional perception and positioning of the working environment,autonomous navigation,autonomous identification of the cage information in the digital pigeon cage,and accurate feeding,to achieve autonomous feeding operations in the loft environment,with an average feeding success rate of 92%. The pigeon precision feeding robot is expected to replace the manual feeding operation,which can effectively reduce the labor demand of the industry and increase the production capacity of the pigeon breeding industry.