摘要
本文提出一种新的车灯尺寸定位误差检测方法,利用图像边缘检测技术,提出轮廓跟踪的算法提取车灯图像的边缘,并采用曲线拟合算法进行分析,快速得到亚像素匹配点。实验结果表明,该方法能精确定位车灯图像边缘,提高了运算速度,优于传统的测量方法。
A new method for automated detection of position error of vehicle headlight shape is introduced,by using the image manipulation edge recognition,introducing arithmetic method of outline track to obtain the image edge of vehicle headlight,and analyzing with curve approximation method,and then the sub-pixel matching point is obtained through evaluating its derivative function.The experimental results show that the method can accurately locate the image edge of vehicle headlight shape and enhance the speed of calculation.The method is better than the traditional measurement.
出处
《气象水文海洋仪器》
2006年第3期39-42,共4页
Meteorological,Hydrological and Marine Instruments
关键词
边缘检测
曲线拟合
亚像素识别
粒子群优化
edge detection
curve approximation
sub-pixel recognition
particle swarm optimization