多算法融合的自适应图像增强方法

Adaptive Image Enhancement Based on Multiple Algorithm Fusion

提出一种多算法融合的图像增强方法,用于工程应用中的复杂降质图像的细节特征恢复.该方法汲取了Laplacian变换法、Sobel梯度法、盒状滤波法、非锐化掩蔽法及灰度幂律法等算法的优点,可对模糊图像进行自适应增强.通过拉普拉斯滤波器和梯度滤波器将原始图像分为基础层、细节层及边缘特征层;对微小细节信息及边缘特征信息进行增强,对基础信息进行压缩;然后采用盒装滤波器对图像的三个分层进行平滑过度及噪音过滤,最后使用非锐化掩蔽法和灰度变换来增加图像灰度的动态范围,从而得到增强后的图像.在相同的工况下,该方法分别与直方图均衡法、自适应伽马矫正法及小波变换的图像增强法实验结果进行对比,结果表明,该方法将图像的清晰度提高了13.1%-126.1%,能有效地处理复杂型感染的图像,避免图像过度增强,可以获得适合人眼的最佳视觉细节内容的增强效果.

For the detail features of the complex degraded images being effectively restored in engineering application, an image enhancement method of the multiple algorithm-fusion was introduced. This fusion algorithm is based on the theory of digital image and integrates the advantages of the Laplace transform, Sobel gradient, box filtering, unsharp masking filter method and gray exponential law strength into the algorithm to enhance the fuzzy images adaptively. The original image, firstly, is decomposed into a base layer, a detail layer and a edge character layer by the Laplace filter and gradient filter. Secondly, the tiny details and edge characteristic information are enhanced and base information is compressed. Then the three layers of the image are processed smoothly and the noise is filtered by the box filtering. Finally, the dynamic range of gray level image is increased by the gray-scale transformation and the unsharp masking method, and the enhanced image is obtained. And under the same load conditons, the proposed method is compared respectively with the traditional algorithms of the HE, AGCWD and WT. The experiment results show that this method can effectively handle the complex images with infection, and the sharpness of the image is increased by 13. 1% - 126. 1%, and avoid the phenomenon of excessive image enhancement, and obtain a superior subjective visual detail effects.