论文部分内容阅读
由于传统蚁群算法搜索空间大,算法时间复杂度高等,导致基于传统蚁群算法的高光谱数据波段选择算法(ACA-BS)耗时长,算法效率低下,且易陷入局部最优。而多态蚁群算法能大大缩小算法的搜索空间,降低算法时间复杂度。因此,研究设计了基于多态蚁群算法的高光谱数据波段选择算法(PACA-BS)。从算法运行时间、波段子集的类别可分性及信息量、总体分类精度等方面对算法进行对比分析。用于实验的数据为Hyperion和AVIRIS高光谱影像。实验结果表明:PACA-BS的运行时间较ACA-BS大大减少;对Hyperion影像进行降维时,基于PACA-BS的运行时间约为ACA-BS的一半。两种算法获得的波段子集的类别可分性大小较为接近,但PACA-BS获得的波段子集的信息量和总体分类精度优于ACA-BS。研究表明PACA-BS是一种效率较高的高光谱波段选择算法。
Due to the large search space and high time complexity of the traditional ant colony algorithm, the ACA-BS based on the traditional ant colony algorithm takes longer time, the algorithm is inefficient and easy to fall into the local optimum. The polymorphic ant colony algorithm can greatly reduce the search space of the algorithm and reduce the time complexity of the algorithm. Therefore, a hyperspectral data band selection algorithm (PACA-BS) based on the multi-state ant colony algorithm is researched and designed. The algorithms are compared and analyzed in terms of algorithm running time, the class separability of the band subsets and the amount of information, the overall classification accuracy. The data used for the experiments are Hyperion and AVIRIS hyperspectral images. The experimental results show that the running time of PACA-BS is greatly reduced compared with that of ACA-BS. When the dimension of the Hyperion image is reduced, the running time based on PACA-BS is about half that of ACA-BS. The class separability of the band subsets obtained by the two algorithms is relatively close, but the information volume and the overall classification accuracy of the band subsets obtained by the PACA-BS are better than that of the ACA-BS. Research shows that PACA-BS is a more efficient hyperspectral band selection algorithm.