一种仿水下生物视觉的大坝裂缝图像增强算法

A dam crack image enhancement algorithm based on underwater biological vision

针对水下大坝裂缝图像非均匀亮度、低信噪比(SNR)和低对比度等特点,提出了一种仿水下生物视觉的大坝裂缝图像增强算法。算法借鉴生物视觉亮度调节特性改善裂缝图像的亮度非均匀问题,并在模拟水下生物"鲎鱼"视觉的侧抑制增强机制机理的基础上,引入自适应的非对称窄条引导模型,对裂缝图像中的线性特征进行增强。理论和实验结果表明,本文算法能够在有效抑制噪声的同时,对图像线性特征增强。

In view of the characteristics such as non-uniform brightness,low s ignal to noise ratio as well as the low contrast of the underwater dam crack image,this paper brings up a novel dam crack image enhancement algorithm,which adopts the simulation of underwater biological vision.With ref erence to the brightness adjustment characteristics of the biological vision,this algorithm also improve s the non-uniform brightness of underwater dam crack image.Furthermore,in order to improve the low signal-to -noise ratio and solve the problem of low contrast of dam crack image,on the basis of lateral inhibition enhancement mechanism of the ＂horseshoe crab fish＂,we introduce the adaptive asymmetric narrow strip guidance models,w hich can help to enhance the linear characteristics of the crack image.The theoretical and the experimental results got from this paper show that the proposed algorithm can significantly eliminate the noises of the underw ater image,and also better improve the definition of the image from the physical standpoint.And at the sam e time,by strengthening the edges of crack image and enhancing the subtle liner structures of interest in th e crack image,this algorithm improves the contrast of the interesting areas,whic h is of great significance for the subsequent crack feature extraction.