The Image Feature Extraction Based On Improved LOG And VG Reconstruction Technology Research

With the development of digital technology, the digital image has become the main way of obtaining the information. But transmitting the high resolution images has been limited by various conditions, In the case of reserving the visual image, the thesis mainly introduce the image feature extraction and vector graphic (VG) reconstruction technology and promote the widespread application of high resolution raster image.As the noise may affect the image quality and the following-up experiment result. The thesis is able to preprocess the image and make the color image gray first, then improving the median filtering algorithm to remove the salt and pepper noise.Feature extraction is the key step of image processing. For the shortcomings of Laplace operator, LoG operator is obtained by Gauss filtering function. Meanwhile the relationship between the Gaussian coefficientδand template width is verified by different operators, and by using different Algorithm to get image threshold and the more reasonable global threshold.The Vector and Reconstruction of Raster Image is the effective strategy to achieve the image transmission. Image boundaries gotten by the calculation above is vectorized to get vector graphics. In the process of vector, the image is triangulated by improved Delaunay Algorithm, which can get the feature point of image and the color of characteristic region to prepare for the reconstruction.Finally, by using SVG file to save, transmit and read the vector feature information, which could achieve and display the graphic transmission in network environment.The experiment result shows that the feature extraction based on LoG and vector based on Delaunay could effectively achieve the network transmission of High resolution images in the case of un-loss of visual information.