关键词: 矿井视频图像, 提升小波变换, 直方图规定化, 双直方图均衡化

Abstract: Undermine illumination is uneven and the surfaces of surveillance cameras are covered by a large number of dust,so,the contrast of obtained video surveillance image is low and there are a large number of granular noise random distributed in video surveillance image.In order to improve the contrast of this kind of image and make full of the effectiveness of the underground video surveillance system,based on lifting wavelet transform (LWT),a adaptive enhancement method of underground uneven illumination image algorithm is proposed.Firstly,the histogram specification (HS) algorithm is adopted to conduct preliminary enhancement of the underground uneven illumination image;secondly,the preliminary enhancement image is conducted lifting wavelet transform,the low-frequency wavelet decomposition coefficient,based on the characteristics of them,the low-frequency wavelet decomposition coefficient is remained unchanged,a new wavelet thresholding function model based on the arcsine function is put forward to filter out the noise distributed in the high-frequency wavelet decomposition coefficients;then,the low-frequency wavelet decomposition coefficients and the filtered high-frequency wavelet decomposition coefficients are conducted refactoring,the underground image without noise is obtained;finally,the bi-histogram equalization (BHE) algorithm is used to improve the visual effects of the filtered underground image without noise.The performances of the algorithm proposed in this paper,histogram specification,counter-peaked mask and wavelet thresholding method are analyzed,besides that,peak signal noise to ratio (PSNR),root mean square error (RMSE) and edge protection index (EPI) are adopted to evaluate the preference of the above algorithms,the results show that the processing effects of the algorithm proposed in this paper is superior to the other algorithms,if has some reference for processing the undermine video surveillance image.

Key words: Underground video surveillance image, Lifting wavelet transform, Histogram specification, Wavelet thresholding function denoising model, Bi-histogram equalization