论文部分内容阅读
针对高分辨率遥感影像进行树冠提取时所遇到的由各类地物之间的相关性和复杂性带来的地物提取难点,将独立分量分析算法和尺度优化法结合进行树冠提取研究.首先,通过独立分量分析算法优化高分辨率遥感影像,去除地物波谱信息之间的相关性,并将ICA变换得到的特征值作为波段加权的权重;再通过改进的最优尺度计算模型选择最优的分割尺度;最后通过对树冠提取的平均精度评价该改进的计算模型.将本研究方法与单纯尺度优化法实验对比分析,结果表明:本文方法有利于降低“同谱异物”和“同物异谱”以及树冠连冠现象,提高树冠信息提取的精度,并可有效避免人为确定分割尺度的主观性和低效性.
Aiming at the difficulty of ground object extraction brought by the correlation and complexity of various kinds of ground objects encountered in the extraction of high resolution remote sensing image, the independent component analysis algorithm and scale optimization method were combined to study the canopy extraction. Firstly, the independent component analysis algorithm is used to optimize the high-resolution remote sensing image to remove the correlation between the spectral information of the features, and the eigenvalues obtained by the ICA transform are used as the weights of the band. Then, the optimal selection model And the segmentation scale is optimized.At last, the improved computational model is evaluated by the average accuracy of tree crown extraction.Comparison of the proposed method and the simple scale optimization method shows that the method proposed in this paper is beneficial to reduce the “Same matter different spectrum” and the crown even crown phenomenon, improve the accuracy of crown information extraction, and can effectively avoid the subjective and inefficient to determine the segmentation scale.