基于BP神经网络高光谱图像分类研究

遥感影像常常存在"异物同谱"现象,影响了遥感影像的分类精度。为了提高分类精度,本文提出了基于BP神经网络的分类算法。采用环境一号卫星HJ-1A星上搭载的超光谱成像仪(HSI)获取的高光谱数据,利用BP神经网络对黄岛区进行遥感图像分类,根据得到的分类结果对原图像进行"异物同谱"现象纠正后重新选取训练样本,然后利用BP神经网络再分类,从而有效解决了"异物同谱"现象。实验结果表明,经处理后的高光谱影像的分类精度得到显著提高,分类总体精度为92.386 5%,比异物同谱纠正前提高了7.83%,Kappa系数也从0.768 2提升到了0.885 8。

Studies on Classification of Hyperspectral Image Based on BP Neural Network

Phenomenon of foreign body with spectrum in remote sensing image has limited the precision of classification.In order to improve the classification accuracy,putting forward algorithm of classification based on BP neural network.The classification of remote sensing image in district of Huangdao was executed by using BP neural network with HJ-1A/HSI data.According to the results of classification were obtained,the spectrum of the original image was corrected and then selecting training samples again,then reclassification of BP neural network solved the phenomenon of foreign body with spectrum to improve the classification accuracy significantly.The results showed that the overall accuracy of the hyperspectral image after processing was 92.386 5%,the increase of about 7.83% than before.The Kappa coefficient was increased from 0.768 2 to 0.885 8.