Research And Implementation Of The Key Technologies In Image Database Retrieval Based On Natural Color Night Vision System

Color night vision technology is one of the key technologies in night vision field.As a frontier algorithm of color night vision technologies,the night vision processing method based on the natural color transferring theory has provided a brand new research method of color night vision technology.With its capacity of endowing the night vision image with natural color,this new method is winning broad attention and being widely studied by lots of scholars.Before color transferring,we need to find out the appropriate natural color image to be the source image to colorize a night vision image.Initially,the selecting of source image of the color night vision system was done by manual work which took quite a long time with poor accuracy and restricted the automation efficiency.To solve this problem,by doing research and implementation of the key technologies of image retrieval technology with the background of natural color night vision technologies,this paper has paid much attention to image retrieval technologies based on the low level image features and has proposed an algorithm of combining Gabor filters and gray-gradient concurrence matrix.Finally this paper has designed and fulfilled an image retrieval system based on natural color night vision technology.The goal of this paper is to automatically find out the source natural color image,which will bring a remarkable breakthrough to the image retrieval technology.The main research of this paper is shown below:Firstly,this paper generally introduces the Content-based Image Retrieval technologies and its development worldwide and then discourses upon the key technologies of image retrieval based natural color night vision system,including image feature extraction,feature similarity measure,image database and relevance feedback technologies. Secondly,this paper studies the theory and several methods of image texture feature which can best describe the characteristics of night vision image.Then this paper proposes a night vision image feature extraction algorithm of combining Gabor filters and gray-gradient concurrence matrix.Featured by its better performance in expressing the texture primitive distribution information of images,gray-gradient concurrence matrix is joined in this algorithm by replacing the traditional way of using gray concurrence matrix.Considering Gabor filters with its excellent behavior in frequency and orientation of partial region of images,the algorithm which combines Gabor filters and gray-gradient concurrence matrix has remarkably enhanced the automation efficiency in the selection of source image and is much more plausible than the other two methods mentioned.Finally,this paper designs and carries out an image retrieval system based on natural color night vision technology by using Visual C++.Adopting SQL Server 2000 as the tool to develop the image database,this paper has fulfilled an image retrieval system with the combination of VC++ and MATLAB,through the utilization of the algorithm proposed by this paper.Then by conducting some experiments and the use of the average retrieval rate,this paper proves the good robustness and image precision of this algorithm and system.