Algorithm Research On Image Edge Detection Based On Wavelet Package

As a newly invented theory, wavelet theory developed rapidly in the last two decades. It has great future in various applications because of its inclusive theory in mathmatics. It had been usded as a common tool in many fields and applications, such as pattern recognition, signal processing, digital television, image analysis, biomedical imaging, theoretical mathematics, fractal, radar and so on.Edge detection had been the focus of theoretical research at home and abroad for a long time. In practical application, the edge is one of the important basic features of one image. In some higher level processing, such as the the image segmentation and image compression, edge detection is a basic tool that leads to further processing on one image.This paper describes several edge-detection algorithms which is based on gradient-operator, such as Roberts, Sobel, etc. The key idea of these operators is that assuming the image-edge-points should be the optimal point of the gray-level-gradient. Then the log-operator and Canny-operator are introduced. These algorithms are sensitive to the noise that the results tend to be relatively poor when the image contains noise. In addition, we also introduce the wavelet-based edge detection algorithm, as well as its general principles and calculation steps. Finally, we introduce a variety of new mathematical tools in the application of edge detection.Based on the wavelet-analysis, this paper presents the image processing with wavelet-packet-analysis. And then, the approximate part of an image is reconstructed to get the results. Then we use the traditional edge detection methods for image processing, we can detect some edge that can not be detected in the original image processing. Contrast with the traditional edge detection methods, it can be seen that the proposed approach has some advantages in restrain noise. This method is suitable for the detection of some weak edges. In the end, we discuss the choice of wavelet basis selection algorithm and the traditional edges detection, as well as the choice of decomposition level. Finally, the license plate location technique is introduced, which includes image pre-processing and location technology on image. And the new edge detection algorithm is applied to it to verify the practicality of the new algorithm. We also introduce some minor remain-drawbacks and the bright development of image detection.