子波变换在红外目标图像边缘提取中的应用

Applications of Wavelet Transform in Infrared Target Image Edge Detection

对红外目标图像进行了图像增强和边缘提取。首先采用直方图均衡、非线性灰度变换、滤波的方法对红外目标图像进行增强、抑制噪声 ,然后基于子波变换的原理 ,采用双阈值自动门限化方法 ,对增强后图像进行边缘提取。分别采用了经典的边缘检测方法、正交子波和非正交二次样条子波对红外图像进行了边缘提取试验 ,结果表明 ,采用非正交二次样条子波提取到的红外图像边缘比采用正交子波及经典的图像边缘检测方法的效果好。提出的方法不仅有较强的噪声抑制能力 ,而且检测到的边缘清晰准确。

A new method to enhance an infrared target image and detect its edge is presented in this paper. It enhances the infrared target image and inhibits the noise by using histogram equalization, nonlinear transform and filtering, then, on this basis of wavelet transform, detects the edge of the enhanced image by using two automatic thresholds. The approaches to extracting image edge with orthonormal wavelet and northonormal quadratic spline wavelet are proposed respectively, then they are compared with the traditional image edge detection methods. Experimental results indicate that the approach to extracting image edge with northonormal quadratic spline wavelet is more efficient than that with orthonormal wavelet and traditional image edge detection methods. The proposed method not only a has strong noise-rejection ability, but also can detect exact and clear edges.