基于机器视觉的车辆保险盒在线检测研究

摘要: 车辆保险盒作为汽车电控系统中的一个重要的元器件，其质量好坏直接影响汽车的性能，传统的车辆保险盒检测主要依靠人工检测，检测费时费力，针对该问题，提出一种基于视觉的车辆保险盒在线检测方法，分析了产品图像校正到标准模板图像的位置误差，采用SURF(Speeded Up Robust Feature)算法和平面单应性理论将待检产品图像变换到标准模板位置，利用颜色直方图匹配和模板匹配完成保险盒上元件的检测。实验结果证明，该方法检测效率高，稳定可靠，能够满足在线检测的要求，具有一定实用价值。

Research on vehicle insurance box online detection based on machine vision

Abstract: Vehicle insurance box as an important component of automobile electronic control system, its quality directly affects the performance of car.The traditional testing of vehicle fuse box mainly relies on the manual detection which is laborious,to solve this problem, this paper presents a method of vehicle fuse box online detection based on vision, which analyzes the position error between the recified production image and the standard template image. Firstly ,the algorithm of SURF and homography are used to transform the production image to a standard template position, then, using color histogram matching and template matching ,the detection is accomplished.The efficiency, stability and application value of this method used for online detection has been proved according to the experimental results.