Research On Image Contrast Enhancement Technology

Image contrast enhancement plays a significant role in image processing,pattern recognition,computer vision and so on.To enhance the complicated high dynamic scene and solve the problem of noise in enhancement,this thesis systematically studies the adaptive local contrast enhancement technology,the combing of noise-depressing contrast enhancement technology and at last have achieved the enhancement algorithm based on CUDA.The result of experiments prove that the proposed algorithms are effective.The main work and achievements of this thesis can be summarized as follows:(1)To solve the problem that Double Plateauts Histogram Equalization(DPHE)algorithm changed the shape of original Histogram,this thesis proposed a new Plateauts Histogram Equalization algorithm based on shape-preserving.Compared with DPHE,the proposed method almost preserved the shape of original histogram while depressed large background information,and this improved details in image.(2)Proposed a new enhancement algorithm based on noise-depressing.The input noise image is decomposed into several layers through guide filter,among them the last layer is base layer and the former layer subtracts later layer produce several detail layers.Then enhance the contrast of the base layer and remove noise in each detail layers.At last,fuse the base layer and detail layers.And this algorithm realized the improvement of contrast and denoising at the same time.(3)In order to enhance the complicated high dynamic scene,this thesis proposed adaptive local gamma correction based on mean value adjustment.In this algorithm,the local mean value and statistic information is regarded as the key factor to determine the gamma value,which could adaptively improve the contrast both in dark and bright areas.Beside,the proposed algorithm have good performance in color resume.(4)The image enhancement parallel algorithms are designed and realized based on CUDA on computers which have NVIDIA graphics card.