| The image edge detection is the base of image further processing and recognition. Reflected in the basic element, the image edge is on the discontinuity point and the rapid change of gray-scale, although the image edge for different reasons. This paper proposes an edge detection method based on the combination of wavelet packet and mathematical morphology, and meanwhile, applied in the edge detection of image"westconcordorthophoto",licence plate character and fire image. The work completed in this paper is as follows:Firstly, the current situations of digital image processing, extraction of licence plate, wavelet and mathematical morphology, digital image processing method and some frequently-used edge detection method are introduced and some of them are given.Secondly, the operation, the standard and steps of several common edge detection methods are researched. And the simulation results of real images using every method are showed.Then, wavelet and wavelet packet based on the basic theories of wavelet and wavelet packet and the correlation theory of mathematical morphology are compared and emulated. The spectrum window of the wavelet transformation is much thinner, so that the better time-frequency window or optimal basis of the analyzed image can be bobtained. And the image character is more obvious, which is propitious to further analysis of image. After that the basic operation of mathematical morphology and the application of operations combination are introduced and simulated. Then the application of morphological gradient at edge detection is investigated and simulated.At last, the edge detection results of image"westconcordorthophoto", licence plate character and fire image using the method proposed in this paper are simulated with MATLAB, repectively. Comparing the results using edge detectors with SNR, the better image edge can be obtained using the method proposed in this paper. |
PDF Full Text Request