Research On New Methods Of Depth Image Segmentation

Range image is rather special,which is different from ordinary grayscale image.It has nothing to do with the direction of the source of light,and there is no shadow.Range image can detect the dark side of object which in conventional optical image.It accurately records the three-dimensional range information of the object surface.Range image data can also be accurately converted to geometry,which facilitates three-dimensional object modeling,object recognition as well as location and reduction of three-dimensional scene.Image segmentation is not only one of the main research directions of image analysis,but also an important intermediate link from image processing to image understanding.In range image segmentation,the traditional algorithm mainly includes edge based segmentation algorithm and region based segmentation algorithm.Traditional algorithms do not necessarily apply to all range image segmentation,therefore,we need to improve the traditional algorithm,or integrate into other disciplines of theoretical knowledge,design a new range image segmentation method according to the given range image.This paper is devoted to search for a better method of range image segmentation.By studying traditional segmentation methods,using classic operator to segment gray image and range image at the same time,we found that the effect of traditional range image segmentation methods is not so ideal.Therefore,in this paper we try to introduce the theoretical knowledge of mathematical morphology as a theoretical basis,combined with the traditional segmentation methods,design two kinds of new segmentation method based on morphology.Before the segmentation,the range image preprocessing step is added,and the improved median filter is used to filter the segmented image.Then through the computer simulation experiment,the same range of the image using different segmentation algorithm to compare the experiment.Then,a simple algorithm evaluation system is designed to select the most suitable segmentation algorithm for the range image of the experiment.The experimental results show that the two new segmentation methods based on morphology are superior to the traditional methods in depth image segmentation.But the time consumption of the two new methods,which sacrifice the time to get better segmentation results,is larger than that of the traditional method...