Research On Design And Control Of Automatic Feeding System For Indoor Mud Crab Breeding

With the continuous improvement of living standards,people’s demand for aquaculture products is increasing.At present,the blue crab breeding method in my country is mainly based on manual labor,but the number of products in this method is limited by uncontrollable factors such as the breeding environment,which cannot meet the market demand for blue crabs.Many aquaculture farmers have begun to try new farming models,and put the intelligent farming equipment into actual production.The indoor recirculating crab farming mode is favored by farmers because of its high degree of intensification and easy management.However,in this farming mode,the feeding of bait still relies on manual operation and the degree of automation is low.The feeding amount is determined according to the farming experience of the farmers,which is full of uncertainties.Feeding too much bait will lead to too much residual bait and the hidden danger of water pollution;feeding too little bait will affect the normal growth of blue crabs.In order to solve this problem,this paper studies the automatic feeding system,designs the automatic feeding system for blue crabs,and realizes the precise feeding of bait based on the machine vision recognition technology.The research content of this paper mainly includes the improvement of YOLOv5 recognition algorithm,the research on the estimation method of feeding amount based on machine vision,and the construction of automatic feeding system.Including the following aspects:(1).Combined with the actual breeding environment of the indoor blue crab breeding mode,in order to solve the problem of low recognition accuracy due to small recognition objects and poor light conditions in the detection environment,the YOLOv5 algorithm has been improved:adding a new method for recognizing small targets The detection layer,the fusion attention mechanism module and the improved loss function.(2).Research on estimation method of feeding amount of blue crab based on machine vision.Collect images in the mud crab breeding box,establish a mathematical model of image data and mud crab breeding information,estimate the feeding amount,and provide data support for subsequent automatic feeding.(3).According to the production requirements of the indoor breeding mode,an automatic feeding system with functions such as automatic feeding and information interaction was designed,and the operation programs of each stage were written to realize automatic feeding.At the same time,an interactive interface with the functions of information interaction and command sending is designed.(4).According to the production environment of indoor breeding,an experimental platform for automatic feeding of blue crabs was built,and an experimental prototype of automatic feeding system was made for testing.The mud crabs were divided into automatic feeding experiment group and artificial feeding control group for breeding experiments,and the experimental results were compared.Through the experiment,it is concluded that the feeding efficiency and bait utilization rate of the automatic feeding group are higher than those of the manual feeding group.