论文部分内容阅读
利用MODIS中5个光谱波段上不同云相态的特性,提出了一种基于BP神经网络的云相态检测方法。首先,分析了所选波段上不同云相态的特性,利用5个波段上光谱图像的反射率、亮温值和亮温差值构成4组特征数据作为输入层,隐层和输出层分别采用优化的传输函数。然后,利用3层前馈型BP神经网络对所选波段MODIS数据进行了云相态检测。最后,将两组测试数据用该BP神经网络算法进行云相态检测的结果与相应MOD06云相态数据进行了对比分析,结果表明该方法能很好地识别云相态,检测平均准确率达到86.11%,计算结果与标准结果平均相关性达到0.874的高度相关,且无需在计算前进行复杂的云和晴空分离处理。
Based on the characteristics of different cloud phases in the five spectral bands of MODIS, a phase detection method based on BP neural network was proposed. Firstly, the characteristics of different cloud phase states in the selected bands were analyzed. Four groups of characteristic data were constructed using the spectral reflectance, bright temperature and bright temperature difference in the five bands as the input layer, and the hidden layer and the output layer were optimized Transfer function. Then, the three-layer feedforward BP neural network is used to detect the cloud phase of the selected band MODIS data. Finally, two groups of test data are compared with the corresponding MOD06 cloud phase data by the BP neural network algorithm. The results show that this method can well identify the cloud phase state, and the average detection accuracy reaches 86.11%. The calculated results are highly correlated with the average of the standard results of 0.874. There is no need to perform complicated cloud and clear sky separation before calculation.