Research Of Fog Degraded Image Enhancement Method And DSP Realization

The research of fog degraded images enhancement has very high theoretical and practical application value. For the problem of low definition of foggy images captured by outdoor video surveillance system, from the image enhancement and this physical model restoration two angles, the fog degraded images enhancement are researched deeply in this paper, the major work has been done as follows:A fog degraded images enhancement algorithm based on improved CLAHE is proposed. Firstly, the sub-images histograms are clipped by the clip limit; Secondly, the clipped pixels are redistributed uniformly to each gray level; Finally, the above histograms are adjusted by the Rayleigh distribution function, and then the defogged image is obtained. Experimental results show that the proposed algorithm can limit the noise and improve the problem of detail loss by image defogged by CLAHE.A fog degraded images enhancement algorithm based on Bilinear Interpolation Dynamic Histogram Equalization is proposed. Firstly, the sub-image histograms are divided by the local minimum; Secondly, the output gray level mapping ranges of each sub-histogram are calculated. Finally, the image is processed by histogram equalization and bilinear interpolation, and then the defogged image is obtained. Experimental results show that the proposed algorithm can effectively solve the uniform distribution problem of fog degraded image histogram.A fog degraded images enhancement based on Mean Shift combined with Statistical Region Merging is proposed. The degraded image is divided into a few regions by MSSRM. Considering the thing that the contrast attenuation is different in each region, each region is processed by the dynamic segmentation histogram equalization, and then the defogged image is obtained. Experimental results show that the algorithm can effectively enhance the definition of distant objects.A fog degraded image restoration method based on improved physical model is proposed. Firstly, the sky region of fog degraded image is separated by hill climbing algorithm, and then the sky region is looked as the set of pixels with maximums of depth. Secondly, according to the depth minimum selected by the user, the depth of any pixel can be calculated by the calculation method based on different region; Finally, the scene depth is took into the physical model and the degraded image is restored. Experimental results show that the algorithm enhances the accuracy of the scene depth and the defogging effect is good.A fog degraded video sequences enhancement algorithm based on wavelet packet decomposition is proposed. Firstly, the background and moving object are extracted from fog degraded video sequences; And then they are defogged Respectively by the image enhancement based on wavelet packet decomposition; Finally, the foreground and background image are fused together and the defogged video sequence is obtained. Experimental results show that the algorithm can processed different frequency bands and has better defogged result and real-time.A DSP real-time defogging system is proposed, the system is designed with DSP TMS320DM642 as the hardware core and the fog degraded image enhancement algorithm based on improved CLAHE as the software core. Experimental results show that the system can achieve a better effect, the running time can meet the real-time processing requirement.