Research Of The Image Mosaic Method Based On Corner Detection

With the rapid development of computer technology and digital image processing technology, image mosaic has become a research focus of computer vision and computer graphics. Image mosaic technology was born because of the limitations of capturing hardware. Because images captured by hardware cannot meet the accuracy or viewing requirements, software technique is needed to produce extended view or to improve the precision of image that is image mosaic. Image mosaic makes a panoramic image from a set of images with overlapping parts. Image mosaic technology has been widely used in space exploration, video compression and transmission, virtual reality technology, remote sensing image analysis, medical image processing and other fields. Image mosaic is mainly composed of three parts, image preprocessing, image registration and image fusion. Images registration is the key part. Image registration method based on the characteristics is not easily influenced by factors such as illumination, rotation, and it performs better because it requires fewer characteristics of relatively pixels.Harris algorithm is the extraction of image feature points detection algorithm one of the most widely used features. The algorithm is simple and practical, of high running stability. It has an obvious characteristic of corner detection accuracy is high. But lies in the deficiency of the Harris algorithm, in the process of extracting angular point, will be extracted to noise, influence the accuracy of image registration, and high time complexity of image registration. According to these two points, the thorough research Harris corner detection algorithm, Based on the DBSCAN (Density-based Spatial Clustering of Applications with Noise, a Noise Density Based Spatial Clustering of Applications) of Clustering algorithm, presents a DHarris corner detection method: Harris algorithm was used to extract feature points, use DBSCAN algorithm to Clustering of feature points. After the cluster is marked as noise point removal, the remaining characteristic point is the corner. I design the image mosaic based on the DHarris corner detection method.The basic idea is:(1) with a median filter for two image preprocessing;(2) image registration, the DHarris corner detection method to detect the corresponding corner, corner point of the maximum correlation coefficient was calculated by the NCC, find two match points of the image, except to point with RANSAC method, the transform model of two images;(3) gradually into gradually a weighted fusion method is adopted to image fusion. Experimental results show that using DHarris corner detection method for feature extraction, in addition to some noise points, improve the accuracy of image matching; At the same time to reduce overall to match the Angle of the number of points, improve the efficiency of image registration.