針對邊界及區域影像所做之影像加強

Abstract

在本論文中，我們提出了一個針對邊界及區域影像所做之影像加強。首先 ，我們先將影像經由Sobel處理，藉著取用不同的門檻值得到兩張初步的 邊界線影像。接著利用7x7的邊界線雜訊過濾器、數學形態學以及細線化 ，可以得到兩張較細緻的邊界影像。這兩張影像合成之後，即得所要的邊 界影像。利用獲得的邊界加上分群著色技巧，我們可以產生一個大略的區 域分佈圖。這些影像區域再經由我們所提出的方法將其灰度分佈圖個別均 量化。已均量化的影像加上測出的邊界合成之後，即得最後的輸出結果。 實驗效果證實所提技巧確實可行。 In this paper, we propose an image enhancement system that emphasizes edges and local details. First, we use the Sobel operator to get the rough distribution of the edges of the input image. Two thresholds, called the high-threshold and low- threshold, are used in the Sobel operator to obtain two images. These two images are then processed using a 7x7 filter to remove the noises around the edges, followed by some morphological operations and a thinning process. Two different edge images are then obtained. Through some processing, we can combine the two edge images to get a better edge image. The edge image and a coloring skill based on clustering can then generate a rough region graph. Some local histogram equalization methods are then applied to the detected regions. The obtained image and the edge image are then combined to get the final output. Experimental results are included to prove the feasibility.