Sub-pixel Edge Detection Of Color Images Based On Principal Axis

The Paper Supported by Natural Science Basic Research Plan in Shaanxi Province of China (Program No.2013JM8043).Since the20th century, with the rapid development of micro size measurement technology of sophisticated devices, laser technology, computer technology, and gradually form a new measurement techniques-image processing technique. In the field of image measurement, the most common technology is the image edge detection, it is one of the key technology of digital image processing and pattern recognition, which impacts the effect of feature extraction and pattern matching, therefore, how to resolve the contradiction between the speed and effect of edge detection in the process of detection has been a hot topic of research at home and abroad. So far, a variety of edge location algorithms has been developed, but due to the defects of algorithm itself and the complexity of the algorithm, leading to the existing edge detection algorithm has great limitations.This paper firstly introduces some commonly used edge detection algorithm, and discusses the basic principle of the algorithm in detail, at the same time, summarizes the advantages and disadvantages of various algorithms, on the basis of it, introduces principal axis analysis method, introduces the basic principle of principal axis analysis method and application, and take it to application for the sub-pixel edge detection of color images, thus puts forward the sub-pixel edge detection algorithm of color images based on principal axis. The algorithm integrates the use of principal axis analysis method, adaptive threshold method and sub-pixel edge detection algorithm of gray images. The algorithm can remove the correlation between each color component of color images, at the same time, it can solve the difficulty of automatic selection of threshold and low accuracy. Therefore, it has good universality and adaptability. At the same time, this algorithm. At the same time, when this algorithm achieves the pixel edge of color images, the color image is divided into several sub images according to certain overlapping proportion, the detail information of the image edge is well reserved by using dynamic threshold for each sub image, to a certain extent, inhibit the influence of noise on the positioning results, thus lay a good foundation for the accurate sub-pixel location of color image edge.On the basis of the traditional algorithms, this algorithm improves the deficiency of the traditional algorithm. The experimental results show that, this algorithm is an effective sub-pixel edge detection algorithm for color image, it can not only eliminate noise, but also preserve the integrity of the image edge information, thus it provides an effective means for practical measurement of images.