论文部分内容阅读
对红外热图像进行增强时，常采用基于直方图的图像灰度线性或非线性变换技术以及中值滤波方法，使图像的目标与背景之间的灰度差别放大并且滤除噪声，从而达到增强图像的目的，其缺点是灰度变换技术易放大噪声而中值滤波方法只能滤除高斯白噪声。由于梯度反映了图像灰度之间的差异，而根据二进小波变换可以建立红外热图像的梯度矢量图，通过增强红外热图像梯度的幅度，可以有效地增强图像对比度。根据脉冲噪声和高斯白噪声的多尺度下小波变换特性，抑制噪声点的梯度且增加目标的梯度幅度，能达到既增强图像又抑制多种噪声的目的。
When the infrared thermal image is enhanced, the histogram-based image gray-scale linear or non-linear transformation technology and the median filtering method are often used to enlarge the difference of the gray level between the target and the background of the image and filter the noise so as to achieve the enhancement The purpose of the image, the disadvantage is that gray-scale transform technology easy to amplify noise and median filtering method can only filter Gaussian white noise. Because the gradient reflects the difference between the image gray levels, a gradient vector image of the infrared thermal image can be established according to the binary wavelet transform. The contrast of the image can be effectively enhanced by increasing the gradient of the infrared thermal image. According to the multi-scale wavelet transform characteristics of impulse noise and Gaussian white noise, the gradient of the noise point can be suppressed and the gradient of the target can be increased, so that both the image enhancement and the suppression of multiple noises can be achieved.