多层次自适应空间系数高斯小波图像边缘检测

Multi-hierarchy and Adaptive Space Coefficient Image Edge Detection Based on Gauss Wavelet

为了在有效抑制噪声的同时,能更准确定位边缘,根据高斯函数平滑图像不因信噪比而异这一特点,提出了一种自适应选择σ(空间系数)的算法。该算法首先利用灰度共生矩阵的惯性特征值来计算适合当前图像的σ值;然后根据该值计算相应的高斯高、低通滤波器,再计算所得低通图像的σ值,并以此类推,直至噪声基本去除;最后将用不同σ值得到的各层次边缘图像按一定准则进行融合来得到单像素宽度的边缘检测结果。实验结果证明,该算法与经典算法、B样条小波算法比较,在去除噪声和准确定位边缘两方面均有提高。信噪比可提高0.47%～6.07%,运算时间增加了0.29%～6.36%。尤其对于分辨率较低的图像(256×256)的边缘检测效果更加明显。

According to the characteristic of Gauss which smoothes the image without reference to SNR, a new algorithm with adaptive or（space coefficient） is put forward to in this paper for purpose of both precise image detection and effective noise restraining. Firstly, inertia of moment of gray level co-occurrence matrix is used to design the σ which is suitable to the current image. Secondly, high-pass filters and low-pass filters are designed according to the σand the next σ is determined out in accordance with the image which is filtered by low-pass filters. The process is repeated till noise is removed basically. At last, images of all levels extracted with different σ are fused to obtain the final image edge with only one pixel wide in accordance by certain rules. Simulation results indicate that when comparing with traditional algorithms and B-spline wavelet, SNR is improved 0.47% - 6.07% , and computing time is increased 0. 29% - 6.36% . The new algorithm proposed in this paper is more efficient in precise image detection and effective noise restraining especially for low-resolution image （ 256 x 256 ）.