针对一阶导数Canny算子边缘检测图像断裂处较多以及连通性较差的问题,研究给出了基于广度优先的边缘连接算法,通过连接边缘检测图像的端点减少边缘检测图像的断裂处数量,在Canny算子的基础上实现了图像边缘的优化.该算法通过调节阈值使得图像端点既能得以较好连接,同时不会增加假边错边的数量.这种基于广度优先的边缘连接算法不依赖于图像梯度和图像自身,可以针对边缘图像断裂处主动做出边缘预测,从而成功连接细小的断裂边缘.同时对优化算法的查全率和连通分量给出了定量评估.研究结果表明,基于广度优先算法的边缘连接算法可以减少Canny算子检测结果中的图像断裂点,提高对Canny算子检测的正确率,有效地改善了边缘图像检测的连通性.%Considering the problem of the fractures and poor connectivity of the image edge detection based on the first derivative Canny operator,we present a method of Canny operator in a given breadth-first algorithm. Here,the edge of the image is optimized on the basis of Canny operator by using of connecting edges to reduce the image edge detection end split. By adjusting the threshold,it is shown that the algorithm can connect the endpoints effectively,while not increasing the number of false edges. It can actively predict the edges,and do not rely on the image gradient and the image itself,thus the tiny fracture edges can be successfully connected. The recall and connected component of optimization algorithm have been quantitatively evaluated at the same time. The research shows that the algorithm can re-duce the image breaking points under the Canny operator detection,enhance the accuracy of the Canny operator and improve the connec-tivity of the edge image.
展开▼
