Correction And Enhancement For Omni-directional Image With A Fisheye Lens

The omni-vision system based on the fisheye lens can get all the information in ahemisphere field of view at one time. The information of the image is stable, not easy to loseand the structure is simple, suitable for building mobile monitoring system.One-time information of the big wide angle of is achieved primarily through absorbing alarge bucket of distortion. Although the information is greatly increased and the scenery ofthe center remains the same, the horizontal and vertical sceneries get distorting and are notsuitable for human to see. If you want to use these serious distorted image information, youneed to correct them to ordinary image. The image correction method based on the principleof perspective projection can solve the problem. According to the fisheye lens and the cameraparameters, we can get the corrected image of any angle.Image quality is one of the important research directions in monitoring field. Image aftercollection, transmission and some other treatments may lose quality inevitably, appearsdistorted and are disadvantaged for people to get reliable information. In order to meet actualneed, we must do some image enhancement processing. Image enhancement is an importantbranch of image enhancement. There are many methods, which are mainly divided into spacedomain and frequency domain. Space domain is operated through the pixel grayscale value.Tthe frequency domain is to operate after the image Fourier transform. This article will firstlyuse the traditional image enhancement method and analyze the advantages and disadvantagesof each method.Wavelet transform is developed on the basis of the Fourier transformation. This articlewill use wavelet transform to do enhancement processing to the image. The imageenhancement based on the wavelet transform are mainly using wavelet multiresolutioncharacteristics. The wavelet transform a image into different scales, the overall image arereflected in the low frequency part, and the detail information are reflected in the highfrequency part. Changing some coefficients in the wavelet transform domain while doing theinverter change can make people process their interested part. The experimental results showthat the image enhancement algorithm based on wavelet decomposition has a very goodimprovement to the treatment effect.