论文部分内容阅读
亚像元定位技术对地表遥感信息提取及农业遥感定量化发展具有重要意义。针对当前国内外亚像元定位研究多集中于亚像元定位模型而缺少模型定位精度影响因素分析的现状,该文开展了定位尺度因素(如重建尺度、影像空间分辨率)和像元空间关系等对农业区域多光谱遥感影像亚像元定位模型精度影响的定量分析。以中国吉林省白城地区洮南市和内蒙古自治区兴安盟突泉县交界农业区为研究区域,以典型的空间引力模型为核心模型,以具有相同光谱分辨率的高分一号(GF-1)卫星8、16 m空间分辨率多光谱遥感影像为基础数据,对重建尺度、影像空间分辨率和像元空间关系等因素对遥感亚像元定位精度的影响进行了探讨。结果表明,对于8 m空间分辨率GF-1遥感影像,当重建尺度为5时,在邻接空间关系下的亚像元定位可达到最佳效果,即由40 m空间分辨率遥感影像重建8 m空间分辨率遥感影像的总体精度为74.67%,Kappa系数为0.604;对于16 m空间分辨率GF-1遥感影像,当重建尺度为4时在象限空间关系下的亚像元定位可达到最佳效果,即由64 m空间分辨率重建16 m空间分辨率遥感影像的总体精度为74.65%,Kappa系数为0.623。此外,重建尺度、影像空间分辨率和像元空间关系对亚像元定位精度具有波动影响,3个因素对应的亚像元定位总体精度最大变幅分别为18.08%、4.39%和0.08%,对应Kappa系数变化最大幅度分别为0.268、0.049和0.006。因此,在不同精度影响因素下,基于空间引力模型的GF-1亚像元定位精度影响因素轻重等级依次为重建尺度>影像空间分辨率>像元空间关系,这可为遥感亚像元定位模型选取、模型参数设置以及适宜的遥感数据选择提供一定参考。
Sub-pixel location technology is of great significance to the extraction of surface remote sensing information and the quantitative development of agricultural remote sensing. Aiming at the current situation that the researches on sub-pixel positioning in the world are mostly focused on the sub-pixel positioning model and the analysis of influencing factors of the model positioning accuracy is lacking, this paper has carried out the research on the influence of the positioning scale factors (such as reconstruction scale, image spatial resolution) And other quantitative analysis of the impact of sub-pixel location model accuracy of agricultural multi-spectral remote sensing images. Taking the agricultural area at the junction of Tunquan County, Baicheng District, Baoding City, Jilin Province, China and Tuquan County, Xing’an League, Inner Mongolia Autonomous Region as the research area, taking the typical space gravity model as the core model and the GF-1 with the same spectral resolution, Satellite 8,16 m spatial resolution multispectral remote sensing imagery as the basic data, the influence of the reconstruction scale, the spatial resolution of the image and the spatial relationship of the pixels on the positioning accuracy of the remote sensing sub-pixel is discussed. The results show that for the GF-1 remote sensing image with 8 m spatial resolution, when the reconstruction scale is 5, the sub-pixel positioning in the adjacent spatial relationship can achieve the best effect, that is, reconstructing the sub-pixel image by 8 m spatial resolution remote sensing imagery, The overall accuracy of the spatial resolution remote sensing image is 74.67% and the Kappa coefficient is 0.604. For the GF-1 remote sensing image with 16 m spatial resolution, the sub-pixel localization under the quadrant space relationship can achieve the best effect when the reconstruction scale is 4 , That is, the overall accuracy of reconstruction of 16 m spatial resolution remote sensing imagery with spatial resolution of 64 m is 74.65% and Kappa coefficient is 0.623. In addition, the reconstruction scale, the spatial resolution of the image and the spatial relationship of the pixels have a fluctuating effect on the positioning accuracy of the sub-pixel. The maximum variation of the overall accuracy of the sub-pixel positioning corresponding to the three factors is respectively 18.08%, 4.39% and 0.08% Kappa coefficient of maximum changes were 0.268,0.049 and 0.006. Therefore, under the influence of different accuracy factors, the grading scale of GF-1 sub-pixel based on the spatial gravitation model is in order of reconstruction scale> spatial resolution of image> spatial relationship of pixels, which can be used for remote sensing sub-pixel localization model Selection, model parameter setting and appropriate selection of remote sensing data provide some reference.