基于通道拉伸和改进MSR的蜗轮副接触斑点图像增强算法

Contact spot image enhancement algorithm of worm gear pair based on channel stretching and improved multi-scale Retinex

在蜗轮副接触斑点视觉检测中,摄像头所采集到的图像整体亮度较低、对比度低,导致图像检测结果不够理想,所以需使用图像增强技术,对图像的亮度、对比度等信息进行增强。为此,提出了一种基于通道拉伸和改进多尺度Retinex(MSR)的图像增强算法。首先,在低光照环境下,对图片进行了通道拉伸,增强了其饱和度和对比度,避免了图像增强过程中出现图像细节颜色失真的情况;然后,为了进一步增加图像细节的对比度,采用改进MSR算法对拉伸后的图像进行了处理,利用自适应权重双边滤波、tanh函数分别替换了传统MSR中的高斯中心环绕函数和对数函数,以减少细节信息的丢失;最后,采用自适应伽马变换对图像亮度做了进一步改善,使图像的质量更高,便于图像的后续处理。实验结果表明:与MSR、基于双边滤波改进的MSR增强算法相比,基于通道拉伸和改进多尺度Retinex的蜗轮副接触斑点图像增强算法的图像增强效果更好;进行图像增强后,图片的均值、标准差、信息熵、对比度相比于原图分别提高了48.3%、61.6%、7.1%、91.3%。研究结果表明:采用该图像增强算法之后,图像的质量和特征信息更好,有利于图像的检测和处理。

In the visual detection of contact spots on worm gear pairs,the overall brightness and contrast of the images collected by the camera are low,resulting in unsatisfactory image detection results,so it is necessary to use image enhancement technology to enhance the brightness,contrast,and other information of the image.Therefore,an image quality enhancement algorithm based on channel stretching and improved multi-scale Retinex(MSR)was proposed.Firstly,in low light environment,channel stretching was carried out to enhance the image saturation and contrast,to avoid color distortion in the process of image detail enhancement.Then,in order to further increase the contrast of the image details,the improved MSR algorithm was used to process the stretched image.The Gaussian center surround function and logarithm function in the traditional MSR were replaced by adaptive weight bilateral filtering and tanh function to reduce the loss of detail information.Finally,adaptive gamma transform was used to further improve the image brightness,so that the quality of the image was higher and the image was convenient for subsequent processing.The experimental results show that comparing with MSR and the improved MSR enhancement algorithm based on bilateral filtering,the enhancement effect of the image enhancement algorithm for contact spots of worm gear pairs based on channel stretching and improved multiscale Retinex is better.Comparing with the original image,the mean value,standard deviation,information entropy and contrast of the image after the enhancement are respectively increased by 48.3%,61.6%,7.1%and 91.3%.The results show that the enhanced image quality and feature information are better after adopting this image enhancement algorithm,which is conducive to image detection and processing.