Study On Algorithm Of Contrast Enhancement For The Large-size TFT-LCD Video Image

Image contrast enhancement engine is the hard core of image processing in the LCD image display, it can enhance the contrast of various kinds of input images, solving the problem of low quality of the images because of the outside impact, such as the brightness, the exposure time and the noise. It is based on the VESA standard with which image contrast enhancement engine is designed in this paper, it can be used in the LCD with the VGA interface, exporting the processed images to the LCD after dealing with the various kinds of images. And it can be treatd as IP core and used in other image display equipment.Traditional image contrast enhancement algorithm are introduced and analyed in this paper, for example, linear convertion, nonlinear convertion, wavelet convertion, PCNN and Retinex theory. Using the software of Matlab 6.5 to simulate the contrast enhancement effect of the processed images with different algorithm after programming, and compare these stimulated images from the point of subject of human vision and objective evaluation.There are some disvantages in the algorithm based on the Retinex theory used widely. Just like processing directly on components of red, green and blue, involving a lot of index operation and log operation with the processing core of Guassian convolution function. Corresponding improving methods are listed in this paper, and compare and analyze the existing algorithm with the improved algorithm after Matlab 6.5 has simulated the images.To the principle of simplifing the hardware circuits and improving the quality of the output images, proportional image contrast enhancement is chosen to be the algorithm used in this design. The design of the structure of the engine of the contrast enhancement, the partition of every functional module are all included in order to optimize the hardware implemention of the contrast enhancement processing module. Functional simulation and timing simulation are completed through the platform of Matlab, QuartusⅡand Modelsim, getting the waveform of the core module, so as to fully test and verify the timing feasibility of this algorithm and the correctness of the coding design.