基于改进局部二值模式算子与蚁群优化的图像边缘提取算法

设计了一种改进LBP耦合蚁群优化的图像边缘提取方案.首先,为了降低对噪声的敏感度,通过一个大小为3*3窗口的均值滤波对输入图像进行处理;其次,在LBP中嵌入一个平滑函数S,构建了一种改进的LBP模型;然后,计算图像的改进LBP值与其权重,通过改进LBP值与权重来体现像素间的差异程度,有效表征边缘的细微变化;最后,引入蚁群优化,通过路径优化与正负反馈求取边缘最优值,提高边缘的清晰度和细节表示.实验结果表明:相对于当前常用的边缘提取方法,本文算法提取边缘具有更高的清晰度和连续性,能减少噪音和模糊干扰,有效识别微小细节.

Edge Extraction Algorithm Based on Improved LBP Coupled ant Colony Optimization

In order to solve the defects which is difficult to both noise elimination and edge preserving in the traditional LBP model, and further improve the clarity and detail information of edge extraction, a image edge extraction scheme has been designed in this paper based on improved LBP and ant colony optimization. Firstly, in order to reduce the sensitivity to noise, the input image was processed by means of a mean filter of size 3*3. Secondly, an improved LBP model was constructed by embedding a smooth function S into LBP. And the degree of difference between pixels was reflected by improving the LBP value and weight to effectively characterization of subtle changes in edges. Finally, the best edge value was got by path optimization and positive and negative feedback, as well as introducing the ant colony optimization to improve the sharpness and detail representation of edges. Experimental results show that this algorithm has higher extraction edge sharpness and edge quality coefficient which can better reduce the noise and fuzzy interference for effectively identifying small details compared with the commonly used edge extraction methods.