基于偏移场矫正的图像断面边缘识别优化算法

Image Edge Recognition Optimization Algorithm Based on Bias Field Correction Section

对图像的断面进行偏移场矫正,增强对正确目标边缘的识别能力。由于图像中存在的灰度不均匀断面特征和线性偏移场,对图像断面的边缘有效识别困难,传统方法采用轮廓线提取方法进行断面边缘识别,受到边缘轮廓线的线性偏移限制,识别效果不好。提出一种基于偏移场矫正的图像断面边缘识别优化算法。通过核函数引入局部灰度信息建立了偏移场矫正模型,将去噪后的断面信号进行灰度均衡预处理,对于相邻区域传递的消息进行收敛性判断,增强了演化曲线对正确目标边缘的识别能力。研究得出,采用该算法进行图像边缘识别,具有更低的差错率。算法将在远程图像识别和地理遥感特征提取探测等领域具有较好的应用价值。

The image of the section of bias field correction, enhancement of the right edge of target recognition ability. Be-cause the gray scale images in the presence of non-uniform section features and linear displacement field, on the edge of the effective identification difficult image section, traditional methods of extracting method of section edge recognition us-ing contour line, contour offset by linear constraints, the recognition effect is not good. A bias field correction section optimi-zation algorithm based on image edge recognition is proposed. The kernel function by introducing a local gray information established correction model displacement field, will carry on the pretreatment to the gray balance section signal noise, for the adjacent zone transfer message to judge the convergence, enhanced the evolution curve recognition ability to the correct target edge. The results of this study, using the algorithm for image edge detection, it has a lower error rate. The algorithm will have good application value in the field of remote image recognition of remote sensing and geographic feature extrac-tion detection etc.