基于神经网络实现废钢车辆的精准定位

Accurate Positioning of Scrap Vehicle Based on Neural Network

为实现废钢等级识别全过程无人干预,首先需要通过摄像机自动对废钢车辆进行定位。由于废钢卸料点周围都是废钢、废钢车辆里装载的也是废钢。按照传统方法,很难精准的定位废钢车辆。本文首先对传统的基于阈值的图像分割方法、基于边缘的图像分割方法、基于区域的图像分割方法进行分析,然后提出了基于神经网络的废钢车辆定位方法,与传统图像分割方法相比,准确率有明显的提升,能够精准的实现废钢车辆的定位。

In order to realize the whole process of scrap grade identification without human intervention,it is necessary to locate the scrap vehicle automatically through the camera. Because there is scrap around the scrap unloading point,the scrap loaded in the scrap vehicle is also scrap. According to the traditional method,it is difficult to accurately locate the scrap vehicle. This paper first analyzes the traditional image segmentation methods based on threshold,edge and region,and then puts forward the scrap vehicle positioning method based on neural network. Compared with the traditional image segmentation method,the accuracy is significantly improved,which can accurately realize the scrap vehicle positioning.