Visual garbage classification system based on TensorFlow and garbage classification method

Abstract

The invention discloses a visual garbage classification system based on TensorFlow and a garbage classification method. The system comprises a conveying belt, a first photoelectric switch sensor, a garbage tray, a slide rail, a camera, a second photoelectric switch sensor, intelligent garbage cans, an NVIDIA machine learning platform and an STM32 controller. The system applies an AI technology togarbage recognition and applies a deep learning visual classification technology to garbage sorting, acquires garbage images by the camera and adopts a TensorFlow deep learning frame, and greatly improves the garbage recognition accuracy by migration training of a MobileNet SSD. An STM32 single-chip microcomputer processes recognition results, controls the conveying belt to separate garbage and controls the slide rail to accurately put garbage into corresponding intelligent garbage cans, carrying and placement of the garbage are realized, human participation is not needed, the working efficiency and accuracy are improved, and the labor cost is greatly reduced.

Description

technical field [0001] The invention relates to the field of garbage classification, in particular to a TensorFlow-based visual garbage classification system and classification method. Background technique [0002] Garbage treatment mainly adopts three methods: garbage incineration, sanitary landfill, and composting. Compared with landfill treatment, waste incineration has the advantages of small footprint, easy site selection, short processing time, significant reduction, thorough harmlessness, and recyclable waste incineration waste heat, etc., but the combustion exhaust gas and waste residue seriously pollute the environment. Although sanitary landfill controls landfill seepage and landfill gas, it still has the disadvantages of occupying a large area and taking a long time to process. Composting only applies to kitchen waste. None of the three mainstream garbage disposal methods classifies recyclable garbage, which is a huge waste of resources. [0003] There is a lar...

AI Technical Summary

Problems solved by technology

Compared with landfill treatment, waste incineration has the advantages of small footprint, easy site selection, short treatment time, significant reduction, complete harmlessness, and recyclable waste heat from waste incineration, but the combustion exhaust gas and waste residue seriously pollute the environment

Although sanitary landfill controls landfill seepage and landfill gas, it still has the disadvantages of large area and long processing time

Composting only applies to kitchen waste

None of the three mainstream garbage disposal methods classifies recyclable garbage, which results in a huge waste of resources

The document "Peng Xinyun, Li Jiale, Li Wan, Liu Xingzhou, Zhang Chengfa, Lin Xianxin, Ou Jiacheng. Research on Garbage Recognition and Classification Based on SSD Algorithm [J]. Journal of Shaoguan University, 2019, 40(06): 15-20." uses SSD target detection The model has carried out garbage classification, but it has only realized garbage classification, and has not yet realized garbage sorting and placement, so it is not very practical; in addition, the SSD target detection model adopted has a large amount of computation and high hardware requirements, so it is not easy to implement on embedded devices accomplish

Images