Design And Research Of Front-end Error Correction Sorting System For Domestic Garbage

With the rapid growth of global domestic waste production every year,but the serious shortage of garbage classification and treatment capacity,which brings about environmental pollution and waste of resources,it is a new idea to use intelligent garbage sorting methods to classify garbage.This paper studies the problem of error correction identification and sorting of multiple tracks at the front end of domestic waste,and develops a flexible,efficient and low energy consumption intelligent identification and sorting system for domestic waste.The research contents are as follows:(1)Put forward the technical scheme of front end error correction identification and sorting system for domestic waste.For the current source image classification of domestic waste,it is high cost and low efficiency to classify one waste at a time;In the later stage,large-scale equipment concentrated garbage sorting,which caused serious mixed deformation and low resource recovery rate.This paper puts forward the design idea of introducing error correcting garbage sorting system into the front end of domestic garbage disposal,and puts forward the technical route and research scheme.(2)Appearance design and functional structure planning of the front end error correction identification and sorting system for domestic waste.Combined with ergonomics,design psychology and other knowledge,complete the system framework design,size function planning,process planning and use scenario design;Complete the manual drawing of product appearance and internal structure,3D modeling of product,product rendering and animation rendering of system operation principle.(3)Hardware design and assembly of the front end error correction identification and sorting system for domestic waste.The system hardware equipment includes four track garbage delivery entrance,garbage separation device,garbage classification assembly line,garbage sorting mechanical arm,garbage collection container,etc;Complete the material selection and procurement of key equipment such as image acquisition equipment,assembly line and mechanical arm of the system,and finally complete the hardware assembly of the garbage sorting system.(4)Development of target recognition system for front end error correction identification and sorting of domestic waste.Complete the collection,filtering and labeling of garbage target image database.Carry out a variety of target detection algorithm comparison experiments,and finally use YOOv5-S model training;The knowledge map drawing and merging experiment were completed,and the IE image data enhancement and expansion algorithm was added to improve the m AP value of recyclable garbage identification accuracy from 76.61% to 95.19%;And completed the training of wet waste identification model and dry waste track total identification model.(5)Integration of error correction identification and sorting system at the front end of domestic waste.Complete the integration of the recognition model YOLOv5 of recyclable garbage and wet garbage with the mechanical arm;Complete the target positioning,inverse kinematics solution and path planning of the manipulator;Complete the static grab and sorting speed test of recyclable garbage;Design the sorting process,and use raspberry pie and relay to realize the fixed-point shutdown of the assembly line;Finally,the dynamic garbage sorting experiment on the conveyor belt of the dry garbage race track was completed,and the sorting success rate was 89.78%.The integration verification of the front end error correction identification and sorting system of domestic garbage was realized.