论文部分内容阅读
以2个小麦品种,4个氮素水平的裂区实验设计,对垂直垄平面上不同观测角度冬小麦冠层光谱与叶片氮素含量的关系进行研究。结果表明,在选择的10种光谱特征参量中,RVI与冠层叶片氮素含量相关性最高,不同角度RVI中,0°观测角的RVI的相关性最显著。对冠层不同角度RVI及叶片氮素进行多种模型拟合,其中0°的GROWTH模型拟合效果最好,建立的0°模型RMSE为0.3402,预测准确率达到90.3%,比180°模型预测精度高23.7%,比90°高5.5%;对2个小麦品种分别建立0°观测角GROWTH模型,普通蛋白型小麦J411模型RMSE为0.2915,反演准确率达到91.64%,ZY9507模型RMSE比J411高31%,预测准确率低3.3%,说明建立的模型更适合普通蛋白含量小麦品种叶片氮素含量检测。
The relationship between canopy spectral reflectance and nitrogen content of winter wheat under different observation angles was studied with two wheat cultivars and four nitrogen fractures. The results showed that the correlation between RVI and canopy leaf nitrogen content was the highest among the ten selected spectral features. The correlation between RVI at 0 ° observation angle and RVI was the most significant in different angles. The results showed that GROWTH model at 0 ° had the best fitting effect, RMSE of 0 ° model was 0.3402, and the prediction accuracy was 90.3%, which was higher than that predicted by 180 ° model The accuracy was 23.7%, which was 5.5% higher than that of 90 °. GROWTH model of 0 ° observation angle was established for 2 wheat cultivars respectively. The RMSE of J411 was 0.2915 and the accuracy of inversion was 91.64%. The RMSE of ZY9507 was higher than that of J411 31%, and the prediction accuracy is 3.3% lower, indicating that the established model is more suitable for the detection of leaf nitrogen content of common protein content wheat varieties.