Multi-temporal RADARSAT-2 Data Research Articles

Synergistic Use of Multi-Temporal RADARSAT-2 and VENµS Data for Crop Classification Based on 1D Convolutional Neural Network

Annual crop inventory information is important for many agriculture applications and government statistics. The synergistic use of multi-temporal polarimetric synthetic aperture radar (SAR) and available multispectral remote sensing data can reduce the temporal gaps and provide the spectral and polarimetric information of the crops, which is effective for crop classification in areas with frequent cloud interference. The main objectives of this study are to develop a deep learning model to map agricultural areas using multi-temporal full polarimetric SAR and multi-spectral remote sensing data, and to evaluate the influence of different input features on the performance of deep learning methods in crop classification. In this study, a one-dimensional convolutional neural network (Conv1D) was proposed and tested on multi-temporal RADARSAT-2 and VENµS data for crop classification. Compared with the Multi-Layer Perceptron (MLP), Recurrent Neural Network (RNN) and non-deep learning methods including XGBoost, Random Forest (RF), and Support Vector Machina (SVM), the Conv1D performed the best when the multi-temporal RADARSAT-2 data (Pauli decomposition or coherency matrix) and VENµS multispectral data were fused by the Minimum Noise Fraction (MNF) transformation. The Pauli decomposition and coherency matrix gave similar overall accuracy (OA) for Conv1D when fused with the VENµS data by the MNF transformation (OA = 96.65 ± 1.03% and 96.72 ± 0.77%). The MNF transformation improved the OA and F-score for most classes when Conv1D was used. The results reveal that the coherency matrix has a great potential in crop classification and the MNF transformation of multi-temporal RADARSAT-2 and VENµS data can enhance the performance of Conv1D.

Selection of PolSAR Observables for Crop Biophysical Variable Estimation With Global Sensitivity Analysis

The role of global sensitivity analysis (GSA) is to quantify and rank the most influential features for biophysical variable estimation. In this letter, an approximation model, called high-dimensional model representation (HDMR), is utilized to develop a regression method in conjunction with a GSA in the context of determining key input drivers in the estimation of crop biophysical variables from polarimetric synthetic aperture radar data. A multitemporal Radarsat-2 data set is used for the retrieval of three biophysical variables of barley: leaf area index, normalized difference vegetation index, and Biologische Bundesanstalt, Bundessortenamt und CHemische Industrie stage. The HDMR technique is first adopted to estimate a regression model with all available polarimetric features for each biophysical parameter, and sensitivity indices of each feature are then derived to explain the original space with a smaller number of features in which a final regression model is established. To evaluate the applicability of this methodology, root-mean square and coefficient of determination were performed under different amounts of samples. Results highlight that HDMR can be used effectively in biophysical variable estimation for not only reducing computational cost but also for providing a robust regression.

Soil Moisture Retrieval From SAR and Optical Data Using a Combined Model

Remote sensing inversion of vegetation-covered soil moisture is often affected by crop canopy, surface roughness, and other factors. In order to eliminate the scattering influence of vegetation effectively, this paper developed a modified vegetation backscattering model to retrieve vegetation-covered soil moisture based on multi-temporal RADARSAT-2 data and field measurements. This model combined the advantages of optical and radar methods by considering scattering contributions of underlying bare soil and vegetation canopy. Vegetation coverage was used to separate the scattering mechanism of the vegetation from bare soil component in a pixel. In addition, advanced integral equation method was presented to define the scattering of underlying bare soil. PROSAIL optical model was applied to calculate crown water content, which is an important variable associated with the scattering of vegetation canopy. Results demonstrated that the modified model on March 29, 2014 performed better in soil moisture retrieval than that at other growth stages with R 2 of 0.806 and root-mean-square error of 0.043 m3·m−3, respectively. Soil moisture can be effectively retrieved by using the modified model in an agricultural region where the surface type is ranging from relatively sparse to full cover. Overall, the modified model provides an insight into extensive application of vegetation-covered soil moisture retrieval in agricultural regions.

Multi-temporal RADARSAT-2 polarimetric SAR for maize mapping supported by segmentations from high-resolution optical image

Due to its ability to penetrate the cloud, Synthetic Aperture Radar (SAR) has been a great resource for crop mapping. Previous research has verified the applicability of SAR imagery in object-oriented crop classification, however, speckle noise limits the generation of optimal segmentation. This paper proposed an innovative SAR-based maize mapping method supported by optical image, Gaofen-1 PMS, based segmentation, named as parcel-based SAR classification assisted by optical imagery-based segmentation (os-PSC). Polarimetric decomposition was applied to extract polarimetric parameters from multi-temporal RADARSAT-2 data. One Gaofen-1 image was then used for parcel extraction, which was the basic unit for SAR image analysis. The final step was a multi-step classification for final maize mapping including: the potential maize mask extraction, pure/mixed maize parcel division and an integrated maize map production. Results showed that the overall accuracy of the os-PSC method was 89.1%, higher than those of pixel-level classification and SAR-based segmentation methods. The comparison between optical- and SAR-based segmentation demonstrated that optical-based segmentation would be better at representing maize field boundaries than the SAR-based segmentation. Moreover, the parcel- and pixel-level integrated classification will be suitable for many agricultural systems with small landownership where inter-cropping is common. Through integrating advantages of the SAR and optical data, os-PSC shows promising potentials for crop mapping.

Tracking crop phenological development using multi-temporal polarimetric Radarsat-2 data

Information on crop phenological development stages such as emergence, flowering, fruiting, maturing and senescence is essential for crop production surveillance and yield prediction. It has long been related to optical spectral signatures such as the Normalized Difference Vegetation Index (NDVI) or spectral shifts in the red-edge range. In recent years, more efforts have been made to explore the sensitivity of Synthetic Aperture Radar (SAR), particularly polarimetric SAR signatures, to crop biophysical parameters or phenological stages. In this study, phenological metrics of canola (Brassica napus) and spring wheat (Triticum spp.) are related with temporal evolution of polarimetric SAR parameters derived from the C-band RADARSAT-2 full polarimetric SAR data. Both crops are very common in north eastern Ontario, Canada, but have very anatomically different development processes. From multi-temporal RADARSAT-2 data acquired in three consecutive years (2012–2014), significant correlations were observed between a number of SAR polarimetric parameters and the growth parameters of both crops. Strong correlation was observed between plant height and the Alpha angle of the Cloude-Pottier decomposition, with the R2 of 0.91 and 0.66 for canola and wheat, respectively. The R2 increased when the polarimetric parameters were smoothed in the time domain (R2 of 0.98 for canola and 0.88 for wheat). Strong correlation was also observed for the two crops between the effective leaf area index (LAIe) and the Beta angle, and between days-after-seeding (DAS) and a combination of the Alpha and the Beta angles. These findings show that multi-temporal C-band polarimetric SAR parameters could be used for tracking crop phenological development stages.

A modified vegetation backscattering model for leaf area index retrieval from SAR time series

ABSTRACTIn this study, a semi-empirical modified vegetation backscattering model was developed to retrieve leaf area index (LAI) based on multi-temporal Radarsat-2 data and ground observations collected in China. This model combined the contribution of the vegetation and bare soil at the pixel level by adding vegetation coverage and the influence of bare soil on the total backscatter coefficients. Then, a lookup table algorithm was applied to calculate the value of vegetation water content and retrieve the LAI based on the linear relationship between the vegetation water content and LAI. The results indicated that the modified model was effective in evaluating and reproducing the total backscatter coefficients. Meanwhile, the LAI retrieval was well conducted with coefficient of determination (R2) and root mean square error (RMSE) of 89% and 0.19 m2 m−2, respectively. Additionally, this method offers insight into the required application accuracy of LAI retrieval in the agricultural regions.

Simplified adaptive volume scattering model and scattering analysis of crops over agricultural fields using the RADARSAT-2 polarimetric synthetic aperture radar imagery

A simplified adaptive volume scattering model (SAVSM) for RADARSAT-2 polarimetric synthetic aperture radar (PolSAR), based on the n ’th power sine and cosine functions, is developed to characterize the changes in crop phenology within the growing season. A three-component model-based decomposition (TCMD) with SAVSM is also implemented with the non-negative eigenvalue decomposition. Multitemporal RADARSAT-2 data acquired in southwestern Ontario, Canada, are used for validating the proposed TCMD-SAVSM. The TCMD-SAVSM is first compared with the adaptive model-based decomposition proposed by Arii et al. in which ∇ f v is set to 0.01, demonstrating that TCMD-SAVSM not only consumes much less computing time but also the decomposed surface and double-bounce components are also more representative of the land cover reality. Using the percentage of power of the remainder matrix 0.001 as the evaluation criterion, the SAVSM achieved the best performance in corn, soybean and wheat fields compared with other popular volume scattering models (VSMs), with the highest mean and lowest standard deviation. It suggests that the SAVSM is a suitable VSM for describing seasonal changes of crop fields. Last, the decomposed surface, double-bounce and volume scattering components of wheat, soybean and corn at various growth stages are analyzed at the individual crop level, and the results demonstrate that their variations are in concert with the crop’s phenological development.

Estimation of snow density using full-polarimetric Synthetic Aperture Radar (SAR) data

In this paper, a new snow density estimation methodology is proposed for full-polarimetric Synthetic Aperture Radar (SAR) data. The generalized four component polarimetric decomposition with unitary transformation (G4U) based generalized volume parameter is utilized to invert snowpack dielectric constant using the Fresnel transmission coefficients. The snow density is then estimated using an empirical relationship. Six Radarsat-2 fine resolution full-polarimetric C-band datasets were acquired over Himachal Pradesh, India. The near-real time in-situ measurements were collected with the satellite pass to validate the proposed method. The mean absolute error (MAE) of the proposed method is 0.027gcm−3 and the root mean square error (RMSE) is 0.032gcm−3. The snow density variation within a season were also analyzed using multi-temporal Radarsat-2 data.