Using the high spatial resolution remote sensing technology to monitor the growth of crop has been a hot issue in current agricultural remote sensing research. In this study, the data from two homemade satellites, HJ satellite with middle spatial resolution and GF1 with high spatial resolution were fused by means of PCA, Brovey, HPF and Wavelet after atomospheric correction and geometric correction. The fusion effectiveness qualitatively and quantitatively evaluated. Among the four fusion methods HPF fusion image presented richer information and better fidelity;its standard deviation and infor⁃mation entropy were 10.984 3 and 1.468 6, and the cross entropy and correlation coefficient between HPF and original im⁃age were 1.848 5 and 0.370 2. The RVI and NDVI of HPF fusion image were 6.508 1 and 0.713 6, the closest to the meas⁃ured values, with their approximation rates of 87�71% and 98�63%. The accuracy of the HPF fusion image was 98�08%, verified by the rice planting area in the extracted samples. The research suggested that using HPF method to merge HJ image and GF1 image could enhance the infor⁃ mation content and spectral characteristics of the fusion image, favorable for improving the accuracy of rice planting area ex⁃traction in county level.%利用中高空间分辨率遥感技术监测农作物生长已成为当今农业遥感研究的热点问题。本研究选用国产中空间分辨率卫星HJ星和高空间分辨率卫星GF1号遥感影像数据，在分别对其进行大气校正、几何校正等预处理的基础上，选用PCA、Brovey、HPF和Wavelet 4种融合方法对HJ星和GF1号遥感数据进行融合，并对其融合效果进行定性和定量评价。通过计算融合影像的RVI和NDVI，比较其光谱特征的差异性，并对筛选出的最适融合影像进行水稻种植面积的提取与精度验证。结果表明，在4种遥感融合方法中，HPF融合效果最佳，其标准差和信息熵分别为10．9843和1．4686，信息量丰富，与原始影像的交叉熵和相关系数分别为1．8485和0．3702，保真度较好，HPF融合影像RVI和NDVI分别为6．5081和0．7136，与实测值最为接近，近似率分别为87�71％和98�63％，利用水稻种植面积提取样方验证的精度为98�08％。说明，对HJ星和GF1号遥感影像进行HPF融合，可以增强融合影像的信息量和光谱特征，有利于县域级水稻种植面积准确提取。
展开▼
