基于Retinex的X光钢丝绳芯输送带图像校正和增强方法研究

Study on correction and enhancement of X-ray image of steel cord conveyance belt based on Retinex

在矿用钢丝绳芯输送带X光图像采集过程中,由于点光源到线阵探测器的各像素点的距离不同,以及每一像素点的物理特性存在差异,造成图像不均匀。采用Butterworth低通滤波器对X光图像滤波,经过去噪、平滑处理,得到去除噪声后的图像;结合Retinex理论建立了基于基函数的不均匀光照背景估计模型,利用该模型对去除噪声后的图像灰度校正,得到校正后的图像;利用拉伸图像对比度的方法增强校正后的图像对比度,从而去除原始图像的不均匀光照的影响,得到增强后的图像。试验结果表明,该算法能有效校正不均匀X光图像,提高了图像对比度,为进一步矿用输送带的检测和图像处理奠定了基础。

In the process of acquiring X-ray image of mine-used steel cord conveyance belt, since the distance from the point source to each pixel of the linear array detector was different and the physical features of all pixels had difference, uneven image formed. Butterworth low-pass filter was applied to filter X-ray image, after denoising and smoothing, the denoised image was obtained. In combination with Retinex theory, uneven illumination background estimation model was built based on basis function, and the model was applied to correct the gray of the denoised image so as to obtain the corrected image. By stretching the image contrast, the contrast of the corrected image was enhanced, thus the uneven illumination of original image removed so as to obtain the enhanced image. Test results showed the algorithm not effectively corrected uneven X-ray image, but enhanced the image contrast, which offered foundation for further detection and image processing of mine-used conveyance belt.