基于PROSAIL模型的高光谱遥感图像模拟研究

针对高光谱图像数据难以获取的情况，提出了一种基于PROSAIL辐射 传输模型的高光谱图像模拟方法。首先，对PROSAIL模型的输入参数进行了敏感性分析。以 此为基础，针对5个典型生化参数(叶片结构参数N、叶绿素含量Cab、叶片干物质含量Cm、类 胡萝卜素含量Car和叶面积指数LAI)，基于支持向量机构建了一种基于多光谱数据的参数计算模型。其 次，根据得到的典型生化参数图，通过利用PROSAIL模型逐像元计算获得了高光谱模拟图像。 对HJ1A环境卫星的CCD1多光谱图像数据进行了实验，实现了高光谱图像数据模拟。在将等效计算后的模拟结果 与HJ1A CCD1多光谱图像数据进行对比后发现，两幅图像对应波段的 平均结构相似度都超过了95%；除了蓝光波段之外，全图的相对误差均小于10%。结果表明，本文方 法可以很好地实现基于多光谱数据的高光谱图像模拟。

Simulation of Hyperspectral Remote Sensing Image Based on PROSAIL Model

For the case that it is difficult to acquire hyperspectral image data, a hyperspectral image simulation method based on the radiation transfer model PROSAIL is proposed. Firstly, the sensitivity analysis of the input parameters of the PROSAIL model is carried out. On this basis, a parameter calculation model based on a support vector machine is established for five typical biochemical parameters (leaf structure parameter N, chlorophyll content Cab, leaf dry matter content Cm, carotenoid content Car and leaf area index LAI). Secondly, simulated hyperspectral images are obtained through pixel by pixel calculation by using the PROSAIL model according to the typical biochemical parameters. The multispectral image data from CCD1 onboard the environmental satellite HJ1A are used to realize the simulation of hyperspectral image data. After the equivalent simulation results are compared with the multispectral image data, it is found that the average structure similarity of the corresponding bands of two different images is greater than 95%. Except the blue color band, the total relative error is less than 10%. This shows that the proposed method can be used to realize the simulation of hyperspectral images based on multispectral data.