论文部分内容阅读
从中国典型地物标准波谱数据库中收集到3年共16d的地面实验测量数据,进行统计分析,得到了棉花各生长期的主要结构参数的分布范围,使用冠层反射率模型SAILH对棉花全生长期的波谱进行了模拟分析,进而模拟得到不同日期棉花像元对应的CEBERS-02 CCD各波段的反射率．并以新疆生产建设兵团农八师143团农场为实验区,用马氏距离匹配和光谱角度匹配两种方法,实现了CBERS-02 CCD图像棉花像元提取．并在最后分析了辐射定标误差对棉花像元识别的影响．结果表明,基于CEBERS-02 CCD数据,使用光谱角度匹配方法可以得到较好的棉花像元识别结果,能够为以后的棉花长势监测提供有力支持．
The ground-based experimental data from 16 years of 3-year total data collected from the database of typical landmarks in China were collected and statistically analyzed. The distribution of the main structural parameters in each growth period of cotton was obtained. Using the canopy reflectance model SAILH, Period of the spectrum were simulated and analyzed, and then get the date of cotton pixel CEBERS-02 corresponding to the various bands of reflectivity. In order to obtain the CBERS-02 CCD image cotton pixels, the 143 farms of Nong Bashi division of Xinjiang Production and Construction Corps were used as experimental areas. Mahalanobis distance matching and spectral angle matching were used as the experimental areas. Finally, the influence of radiation calibration error on cotton pixel recognition is analyzed. The results show that based on the CEBERS-02 CCD data, using the spectral angle matching method can get better cotton pixel recognition results, which can provide strong support for future cotton growth monitoring.