Study Of Image Segmentation Method Based On Histogram And Wavelet Transform

Image segmentation is an important technology of image processing, the purpose of which separates the object into regions with special character and extracts interesting region. The result of image segmentation will be used to analysis and understand image. It is a classical difficult problem. No method which has good effect for general images has been found since being researched.In this paper, image enhancement methods will be discussed and histogram equalization, smoothing and sharpening methods will be introduced. Then a survey to the image segmentation methods will be given. By researching, it was found that one of the difficult problems in image distinguish actual edges from noise. Because both edges segmentation and noise are high frequency signals, it is very difficult to distinguish them. If we use lowpass to filter noise, the edges will become illegibility. Wavelet transform appeared recently may resolve this problem, which is a sort of time-frequency analyzing method. If wavelet transform is used to detect edges of image, noise will be restrained in large scale and edges will be orient accurately in small scale. So the problem is resolved. The Methods of edge detection based on wavelet transform have good effect.The method of the wavelet combing with histogram is researched. The application of wavelet in traditional image segmentation based on gray histogram of image mainly makes use of the multiresolution character of wavelet. At different scale, we can choose different thresholds, this may make the threshold more reasonable. At last Experimental results show thatthe method makes the segmentation have good effect.