Research On Sorting Equipment For Waste Plastic Bottles Based On Machine Vision

The sorting work is an important process in the recycling process of waste plastic bottles.At present,the sorting of domestic waste plastic bottles is mainly based on manual operation,which has problems such as excessive labor intensity and low sorting efficiency.Based on vision technology,this subject proposes a scheme for automatic identification and sorting of waste plastic bottles,and studies the algorithm models,machinery,hardware,and software involved in the scheme.Based on optical transformation and geometric transformation,seven different color plastic bottle inspection data sets were produced and ex Panded.Through the deep learning target detection algorithm SSD,the recognition and positioning of waste plastic bottle images were studied,according to the plastic bottle inspection data set,model were trained and tested,and establish an SSD algorithm model suitable for the plastic bottle of this subject.Using fluid simulation software to simulate the load on the plastic bottle when it is sprayed by high-pressure gas,in order to better se Parate the plastic bottles moving at high speed on the conveyor belt,according to the obtained load situation,the se Paration movement model of the plastic bottle after spraying is established.Based on the se Paration motion model,the trajectory of the se Paration motion of plastic bottles under different working conditions(jet angle,injection pressure,conveyor speed)is studied,and the relevant working conditions Parameters are optimized.By analyzing the hardware and structural requirements of the plastic bottle sorting equipment,to selecte and design the hardware modules(image acquisition module,image processing module,and valve drive module).At the same time,the mechanical structure of the camera bracket,adjustment of the nozzle spray angle and position was designed.In view of the possible problems of duplicate targets and incomplete bottles in two adjacent frames of the plastic bottle image sequence obtained by the camera on the conveyor belt,a corresponding recognition processing algorithm is proposed.The software for the control system of plastic bottle sorting equipment has been developed,including:camera calibration module,target detection module,image sequence processing module,plastic bottle sorting module,and which also programme and test each of module.A device for automatic identification and sorting of waste plastic bottles was developed,and a related experimental platform was built.The free fall trajectory of plastic bottles without spraying and the trajectory of movement during spraying were studied experimentally,which further verified the theoretical analysis results.