Image Processing And Analysis Of Objects Attached Each Other Or Affected By Asymmetrical Lights

Some images, grabbed from scene, are more or less destroyed by noise, asymmetrical lights or surface shape of some touched objects, so few of images are segmented successfully. Recently more and more people paid attentions to this technology. The research of this paper is focused on some methods how to segment the interested foreground from images and how to split reasonably those objects touched each other. The aims of the research are to detect those objects automatically. In the end of article, the results of image segmentation will be given.At first, this paper analyzes the CCD model grabbing images and the noise structure of the images. We make some contrastive researches on many image pretreatment technologies as well. After the studies of i he GAUSS f i ltering and median filtering, a new median filtering algorithm is presented that can keep the images' clear edges after smoothing the noises quickly.Secondly, the edges detected by some methods are the important information of objects in images, but those edges are generally unrounded, under the situation the segmentation of images is difficult to be made certainly. Partly based on the information of gray images this paper presents a new algorithm that can segment thoseinterested parts from the background after some regions are adapted automatically to rise. The new algorithm can effectively restrain the negative affects of asymmetrical lights. But this algorithm can' t split those objects that attached to each other. So this paper presents an other short-route splitting algorithm(SRSA) of binary objects , which the distance of the joints of the objects is much shorter than that of the interior.At last, we make some experimentation to check the validity of those new algorithms on various aspects. The results indicate that the interested objects are detected successfully and correctly.As we can see, the two new algorithms can effectively segment images that have been destroyed by the noise or asymmetrical lights.