基于图像分析的后视镜失真率自动检测方法  

Automatic Distortion Detection for Automobile Rearview Mirror Based on Image Analysis

作　　者：秦筱楲[1] 丁明跃[2] 

机构地区：[1]华中科技大学图像识别与人工智能研究所,湖北武汉430074 [2]图像信息处理与智能控制教育部重点实验室,湖北武汉430074

出　　处：《计算机工程与科学》2005年第7期57-59,共3页Computer Engineering & Science

摘　　要：传统的机动车后视镜失真率检测方法采用模拟相机拍照、照片冲洗、人工测量等步骤,效率低下且人为因素对测量结果影响较大。本文使用数字相机成像,针对数字图像,设计并实现了基于图像分析的失真率自动检测算法,通过圆心检测、图像分割、直线与圆交点检测等几个步骤,完成了自动检测功能。该方法高效、准确,达到了实用水平。In the distortion detection for automobile rearview mirrors, the traditional method is photographing with an optical camera, developing photographs and measuring artificially. Due to the influence of artificialness, it is inefficient and inexact. This paper introduces a new automatic detection method based on the computer analysis of the images from digital camera. By the steps of circle centre detection, image segmentation and cross detection, the method succeeds in detecting distortion automatically, efficiently and exactly. Experiments show it is effective in actual applications.

关 键 词：后视镜 失真率检测 HOUGH变换 CANNY算子 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]