Forest Height Inversion Based On Three-layer S-RVoG Model Using PolInSAR Image

The vertical structure of forest canopy has always been regarded as an important parameter for measuring forest changes.The measurement of the forest canopy contributes to the monitoring and management of the ecosystem.Nowadays,we can obtain nearly every piece of forest information by means of remote sensing,in which the SAR system has attracted wide attention in forest ecological monitoring.The research on forest area parameter inversion of images is an important branch of SAR field.The backscatter model can be used to obtain forest parameters that are difficult to obtain by other remote sensing methods,especially forest height.At present,the backscattering model is dominated by the RVoG model.However,because the scattering process of the forest region is too complex,errors will occur in the inversion based on the RVoG model for the uneven terrain of the mountain area.It is a major research direction to improve the scattering model in order to increase the accuracy of inversion.This paper mainly focuses on the forest height inversion of Pol In SAR images under the three-layer S-RVoG scattering model.Firstly,the Pol In SAR technology is described,including interference coherence and coherence optimization theory.Afterwards,by thoroughly discussing the RVoG scattering model and analyzing the effects of various parameters on the interference coherence,a three-layer S-RVoG scattering model is deduced which includes the canopy-tree trunk-slope surface by combining the S-RVoG model and the general three-layer scattering model.Three-layer S-RVoG model is divided into simple form and complete form.The new model takes into account the effect of terrain slope.At the same time,according to the scattering characteristics of the trunk layer,the two-layer model is extended to three layers,which is more suitable for the real scattering process and of very helpful for the forest height inversion in the mountainous region.To solve the problem of inaccurate forest parameters inversion method,after analyzing the advantages and disadvantages of the three-stage inversion algorithm and coherent amplitude method,a high precision bybrid inversion method is proposed for RVoG model which combines the three-stage algorithm with phase diversity optimization and the target decomposition method.And for the simple form of the three-layer S-RVoG model,the table lookup process of the three-stage inversion algorithm was modified in order to adapt to the new model;for the complete form of the three-layer S-RVoG model,the scattering mechanism elimination method was introduced to realize the forest parameters estimation.Finally,the forest parameter inversion experiments are implemented by using simulated data and real data respectively,and the results are evaluated from the aspects of parameter inversion accuracy and root mean square error.The experimental results show that for data with different slopes,the parameter inversion based on the three-layer S-RVoG model can better fit the results and have higher inversion compared to the traditional RVoG model.The root mean square error also has a certain decrease.The three-layer S-RVoG model established in this paper is conducive to improve the application of Pol In SAR image forest parameters inversion in mountainous areas.However,since the model contains more parameters,the parameter inversion process is more complicated and usually requires more processing time to get the result s.