摘要
针对传统的随机抽样一致性算法在精确匹配中计算量大、效率低等问题,提出了一种结合改进粒子群算法的RANSAC精确匹配算法。首先,利用微分流形中单位分解的知识将图像分成几个部分。其次,利用改进粒子群算法选择最佳叶节点进行模型参数估算。最后,保留N个最佳叶节点,返回最优模型,统计几个局部的精确匹配点。通过仿真实验与传统的RANSAC和GASAC进行比较发现,结合改进粒子群算法的RANSAC精确匹配方法,在匹配准确率和结果不稳定等方面有很大的提升,减少了错误匹配点数。
A RANSAC accurate matching method based on improved particle swarm optimization algo-rithm is presented in this paper to solve the problems of large calculation amount and low efficiency of accurate match in traditional random sampling consistency algorithm. Firstly, divide images into several parts by using the unit decomposition method in the differential manifold. Secondly, estimate and calculate the model parameters by using the improved particle swarm optimization algorithm to select the best leaf node. Finally, return to the optimal model with the N best leaf nodes kept, and calculate the accurate matching points in the differ- ent parts. By comparing the simulation experiment and the traditional RANSAC and GASAC, it finds out that the RANSAC accurate match method based on the improved particle swarm optimization algorithm has greatly improved the matching accuracy and the results stability, and reduced the number of false matching points.
出处
《机械与电子》
2017年第7期18-22,共5页
Machinery & Electronics
关键词
粒子群算法
RANSAC
单位分解
精确匹配
particle swarm optimization
RANSAC
unit decomposition
accurate match
