Design of plastic sorting method and implementation on Zynq
The recycling and reuse of waste plastics is one of the important methods to solve the current environmental pollution and resource waste caused by plastics.The difficulty lies in the low efficiency and slow speed of using traditional methods to identify and classify plastic types.This article presents a method for plastic sorting using a near-infrared spectrometer to acquire plastic spectral images and a convolutional neural network algorithm based on Zynq.The hardware design and performance test of the system were carried out on Xilinx Zedboard.The system is hardware designed and performance tested on the Xilinx Zedboard development board,and the data is compressed using fixed-point quantization and the data storage method is optimized.Finally,with a forward inference speed of 0.13 ms and a recognition accuracy of 92.6%,the system successfully classifies six types of plastics including polyethylene,polypropylene,and polystyrene.
万方数据
维普
