针对红外图像普遍存在的信噪比低、目标与背景对比度低、图像边缘模糊、边缘检测难等特点，从人眼视觉仿生的角度，提出一种基于人眼微动机理的小波域红外图像边缘检测新方法。首先对红外图像进行静态小波分解，得到图像在小波域的系数矩阵。再对小波系数在水平、垂直、45°斜向、135°斜向共四个方向进行微动，生成小波域综合移动阵序列；计算移动阵序列与原小波系数的差值；引入竞争机制，得到小波域反映图像边缘信息的系数矩阵。最后，设定阈值进行阈值化、细化处理，根据图像静态小波域与空域的对应关系，检测出红外图像的边缘。实验结果表明，该算法在红外图像边缘检测准确性方面具有明显优势。

Infrared image has the characteristics of low signal-to-noise ratio, low contrast between object and background, blurred edge of object and difficult edge detection. A new wavelet domain infrared image edge detection method based on eye microsaccade mechanism is proposed. Firstly, stationary wavelet transform is performed on the original infrared image, thus the wavelet coefficient matrix is obtained. Then, the wavelet coefficient comprehensive movement matrix sequence is generated through small moving the wavelet coefficients along horizontal, vertical, and diagonal directions. Calculating the difference values between comprehensive movement matrix sequence and original wavelet coefficient matrix, then leading into competition mechanism, the wavelet coefficient matrix that reflects image edge information is obtained. Finally, the wavelet coefficients are processed by thresholding and thinning. According to the corresponding relations of stationary wavelet domain and spatial domain of image, the infrared image edge is extracted. The experimental results show that the method proposed in this paper has obvious advantage in detecting infrared image edge accurately.