苗佳 赵永来
摘 要: 提出一种基于灰度像素邻域模板匹配的轿车流线型曲面缺陷的图像识别与检测方法,采用图像处理方法进行轿车流线型曲面缺陷识别,提高轿车表面缺陷的智能修复能力。对采集的轿车表面图像进行降噪提纯处理,利用自适应权重均衡分割方法进行图像分割,采用Retinex信息增强方法进行曲面缺陷图像增强处理,结合灰度像素邻域模板匹配方法进行缺陷识别与检测。仿真结果表明,采用该方法进行轿车流线型曲面缺陷的图像识别,检测的准确匹配概率较高,对缺陷部位的定位较准,性能优越。
关键词: 图像识别; 曲面缺陷; 图像检测; Retinex信息增强
中图分类号: TN911.73?34; TP391 文献标识码: A 文章编号: 1004?373X(2017)20?0095?03
Abstract: For car′s streamline curved surface defect, an image recognition and detection method based on grayscale pixel neighborhood template matching is put forward in this paper. The image processing method is used to recognize the streamline curved surface defect of car to improve the intelligent repair ability of car′s surface defect. The captured surface image of the car is processed with noise reduction and purification. The adaptive weight balanced segmentation method is used for image segmentation. The Retinex information enhancement method is adopted to enhance the image with curved surface defect, and combined with the grayscale pixel neighborhood template matching method to recognize and detect the defect. The simulation results show that the method has high accurate?matching probability for detection, precise positioning for defect parts and high performance while it is used to recognize the streamline curved surface defect of car.
Keywords: image recognition; curved surface defect; image detection; Retinex information enhancement
0 引 言
图像作为数字信息处理的重要应用方向,在机械加工、目标识别、表面缺陷检测和信息修复等领域都展示了重要的应用价值[1]。研究图像处理在轿车表面曲面中的缺陷检测识别中的问题,由于采用人工校准方法进行轿车流线型曲面缺陷识别受到视线和光照等因素的影响,对缺陷部位的准确定位检测准确性不好。因此需要研究一种智能的轿车流线型曲面缺陷的检测方法,提出基于灰度像素邻域模板匹配的轿车流线型曲面缺陷的图像识别与检测方法,结合图像处理方法进行缺陷识别和检测,取得了较好研究成果。
1 轿车流线型曲面图像采集处理
1.1 轿车表面图像采集
为了实现对轿车流线型曲面图像的缺陷部位……