一种基于Hough变换和图像融合的快速破损检测

A Fast Breakage Detection Based on Hough Transform and Image Fusion

针对工业破损检测的现状进行研究,提出运用数字图像处理系统代替传统的硬件检测。首先对薄膜进行图像预处理,运用Hough(霍夫)变换分割出连续2帧的感兴趣区域,运用小波变换增强图像对比度;其次计算各个像素点的EOI(拉普拉斯能量)和VI(对比度),将归一化处理后的值作为图像融合的系数;最后把融合后图像转化到HSI(色度、饱和度、亮度)空间,进行阈值分割,进行破损判别。实验表明,这种检测方法具有抗干扰性强、成本低、准确性高、计算量小、识别速度快的优点,基本能实现包装的实时检测。

The status of industrial breakage detection has been studied , we propose making the digital image processing system replace the traditional hardware detection , then the first is image preprocessing to membrane , using Hough transform to segment region of interest for two consecutive frames , image contrast enhancement using wavelet transform , the followed is to calculate the energy of Laplace（EOI） for each pixel and Visibility（VI）, the normalized value of the processed image fusion as a factor , and finally fused image was transformed into HSI space , then is the threshold segmentation , to discriminate the damaged .Experiments show that this detection method has strong anti-interference, low cost, high accuracy, small amount of calculation to identify the advantages of speed , meets the requirements of the real-time detection in the industrial field .