轨道巡检图像增强方法研究

针对高速铁路综合巡检车拍摄到的轨道巡检原始图像在某些工况下成像质量差的问题,基于视觉系统神经电生理试验提出了RGC的nCRF模型.由于感受野模型中存在不同的高斯通道,通过设计能自适应显示抑制强度变化的势函数改进了基本模型,提出mnCRF模型来改变抑制效果.根据边缘对比度指数QEC、细节增强指数QDE、灰度保持指数QMG...

Research on Image Enhancement Method of Track Inspection

Aiming at the poor imaging quality of the original images of track inspection taken by the comprehensive inspection vehicle of high speed railway under some working conditions,an n CRF model of RGC was proposed based on the neuroelectrophysiological test of visual system.Because there are different Gaussian channels in the receptive field model,the basic model was improved by designing a potential function that can adaptively display the change of suppression intensity,and a mn CRF model was proposed to change the suppression effect.According to the three image enhancement measurement indexes of edge contrast(QEC),detail enhancement(QDE)and maintain gray(QMG),n CRF and mn CRF models were used to strengthen the image of the image processing standard test library and the high speed railway track inspection image collected by the comprehensive inspection vehicle,and the processing results were compared with those of HE and Retinex algorithms.The results show that the n CRF and mn CRF models are better than comparison algorithms,which can enhance the image of small components on the track inspection image,such as STF and TP.At the same time,they have the image processing functions of edge enhancement,detail enhancement and gray maintenance,which is conducive to the subsequent intelligent analysis and processing.