Indrawing on the Picture Edge Based on Wavelet Technology

摘要

Have studied the wave technology'sapplication in drawing on the picture edge, the local segmentation problem preliminarily solved existing in the picture. The traditional chain-length concept is discarded and the percentage concept of the chain is proposed. The edge-growth method is adopted to get the good closed image. The edges checked by wavelet transforms are of single pixel and accurate orientation and wavelet transform on every image scale can see some edge information. An evident method is to integrate all edge images with all scales in an attempt to play their different advantages to gain accurate edge with wide single pixel. Based on peculiarities of wavelet transform, we can adopt a special general arithmetic to obtain image edge with wide single pixel. Taking advantage of wavelet transform technology, we can not only avoid some trivial and uncertainty brought by setting gray threshold but also wavelet has anti-noise function and is able to capture image details. Most importantly, wavelet can be usable in all images with various gray levels. Firstly, we use designed adaptive filter to filter the waves and then transform wavelet for three times, so we can see wave filter maintain edge while getting rid of noises, Lastly the image segmentation is implemented by finding large regions.