Font Size: a A A

Study On The Extraction Method Of The Closed Contour Of The Object Under The Natural Environment

Posted on:2020-02-19Degree:MasterType:Thesis
Country:ChinaCandidate:J H KangFull Text:PDF
GTID:2428330599460558Subject:Engineering
Abstract/Summary:PDF Full Text Request
Vision is one of the most advanced perceptual organs of human,and visual perception is feedback by obtaining image information from the outside world,so image plays an important role in human perception system.The cognition of human image is to obtain the feature information in the image firstly,and the contour of the object is one of the most important features.Only by obtaining the contour can we recognize and locate the image effectively.The highest goal of computer vision research is to use computer vision to simulate the human vision.Obtaining effective information(such as edges,contours,etc.)from images in visual applications is in a link-to-bottom position in the system.On the basis of previous visual mechanism simulation,this paper remodels the non-classical perceptual field in human visual processing mechanism and constructs an algorithm to simulate the human contour closure mechanism,which increases the extraction effect of object contour in natural images.Firstly,the related research of contour detection and contour closure is expounded,the principle of edge detection and the significance of are introduced in detail,and the closed psychological state of objects in human vision is briefly described.It also introduces the graph theory and Gestalt rules which are close to the closed perception of human psychology.Secondly,Considering that the inhibition mechanism of non-classical sensory field in simulated human visual cells is mostly in the static state,ignoring the mechanism that the eyeball is moving in the eyes of the outside world.By studying the important role of eyeball in human vision,a new filter combining micro-movement of eyeball and non-classical sensory field is proposed.Third,When the human is looking at the external object,even if the object is incomplete,the human vision can close the obtained contour and make better recognition.Considering the influence of pseudo-edge on Gestalt completion rule,a new contour closure algorithm is proposed,which combines the above-mentioned filter and the idea of expressing edge segment connection based on graphics.Finally,the above algorithms are tested with standard image database,and compared with other algorithms of the same type.The qualitative analysis and quantitative analysis of the experimental results are carried out.
Keywords/Search Tags:eyeball movement, non-classical feeling field, graph theory, gestalt completion rule, Contour closure
PDF Full Text Request
Related items