Research On Detection And Recognition Method Of Significant Image Target Based On Spatial Domain

In the field of machine vision and automatic target detection and other significant target extraction has a very important role. Machine vision is the use of machines to simulate the human eye detection and recognition to do. All will be taken into the target image signal, and image processing, according to the image color, brightness and pixel distribution information into digital information. Target recognition can be said is very important to perceive the external world the function of biological visual system. The requirement of automatic target detection in the image information read from the implied, such diversity can be shape, color,shape, texture, motion can also be a relationship with each other, the surrounding objects,prior knowledge / expectations that reasoning. So not to take a single feature recognition can be completed, salient object detection is one of key steps.Image saliency is the important image visual features, since 1998 the ITTI method,attracting large number of scholars on the salient object extraction studies were carried out,but most methods or based on gray image development, due to the human eye view first capture the visual stimulus is part of the, and we have to study the image color image salient object extraction is to extract the eye view of the most sensitive place.This paper studies the processing method based on color information of the salient object,study and try the effect of a variety of color descriptors, study on Extraction of the color information, color descriptor is relatively large impact factors, and ultimately selected using 5model of color feature description scheme.Repeated attempts to color information in the image and the Renyi image of the salient features of the image, the proportion of the image, the final effect of a good significant figure,and write code to view the binarization image in this case.Draft above need to be explored and studied, based on the first to understand the distribution of Wigner-Ville, Renyi entropy by adjusting to adjust the angle ofone-dimensional PWVD horizontal and vertical direction, and then the Renyi entropy diagram for processing, so as to obtain color images based on salient target image after the two value of Renyi entropy Wigner-Ville distribution. The color information of then the color image processing, color information processing, through the comparison with Canny edge detection method(step edge detection algorithm is optimal), image feature extraction of color feature information significantly. Finally, the target image features significantly Renyi image entropy of Wigner-Ville distribution and extraction of color information of the salient features of image according to a certain proportion detection of the final synthesis and extraction.The salient object extraction is a basic step of image processing, this paper is a study on the target extraction method was relatively new color image, first color image into gray image using discrete Fourier transform and form extraction gray change information, calculate the Wigner-Ville distribution of space domain / space frequency domain, special research the nature of their joint showing, to calculate the distribution of the pseudo Wigner-Ville(PWVD)the final statistical characteristics of Renyi entropy after extraction combining color information of color image according to a certain proportion of drawing the final binarization image. This method has important significance for realizing automatic detection for color images of complex background a significant goal. In the experiment, the proposed color image salient object extraction method and classical algorithm were compared. Results show that the proposed algorithm has Efficiency.In the information age can be said that people in all aspects of life all cannot do without the information of the image, so that image processing research and development is very important for social life. Better image processing method plays a very important role in our life and the development of science and technology, and caused a lot of research on machine vision industry began to the automatic target detection. It is significant in the military, medical,plays a very important role in transportation and other aspects. And it is the basis of image processing technology, and significantly better image on this topic is also a big research space.Especially in general electronic devices(such as mobile phone and tablet computer. The new)or provide space for the development of machine vision image processing very large environment, people more in-depth study of various aspects of image processing and, This basic research has also been gradually improving.