Integration of Sentinel-1 and Sentinel-2 for Classification and LULC Mapping in the Urban Area of Belém, Eastern Brazilian Amazon

Abstract

In tropical regions, such as in the Amazon, the use of optical sensors is limited by high cloud coverage throughout the year. As an alternative, Synthetic Aperture Radar (SAR) products could be used, alone or in combination with optical images, to monitor tropical areas. In this sense, we aimed to select the best Land Use and Land Cover (LULC) classification approach for tropical regions using Sentinel family products. We choose the city of Belém, Brazil, as the study area. Images of close dates from Sentinel-1 (S-1) and Sentinel-2 (S-2) were selected, preprocessed, segmented, and integrated to develop a machine learning LULC classification through a Random Forest (RF) classifier. We also combined textural image analysis (S-1) and vegetation indexes (S-2). A total of six LULC classifications were made. Results showed that the best overall accuracy (OA) was found for the integration of S-1 and S-2 (91.07%) data, followed by S-2 only (89.53%), and S-2 with radiometric indexes (89.45%). The worse result was for S-1 data only (56.01). For our analysis the integration of optical products in the stacking increased de OA in all classifications. However, we suggest the development of more investigations with S-1 products due to its importance for tropical regions.

Predicting Forest Cover in Distinct Ecosystems: The Potential of Multi-Source Sentinel-1 and -2 Data Fusion

“…This study further analyzed the impact of sensor fusion of Sentinel-1 and Sentinel-2 to improve forest monitoring in highly variable ecosystems. Similarly to findings of forest monitoring related studies while using optical and radar satellites, the addition of Sentinel-1 data did not significantly improve the overall classification accuracy [26,28]. In both studies, differences between the accuracy of a Sentinel-2 only and fused classification (Sentinel-1 and Sentinel-2) ranged between 1 to 3 %.…”

Urban Ecosystem Services Quantification through Remote Sensing Approach: A Systematic Review

“…Machine learning (ML) techniques, such as random forest (RF), support vector machine (SVM), and artificial neural networks (ANN) are tended towards for the LULC classification and identification of ecological variables. The accuracy of the methods is increasing along with the diversity in its modes of application, this is because of the popularization of the techniques [41,42,[106][107][108][109], even for UES studies [9,69,81,82].…”

“…However, current RS methods can vary greatly. For instance, different applications include the LULC change detection [39] and forest disturbance history [40], data fusion of optical and radar data for precisely machine learning supervised mapping of LULC [41,42], the evaluation of water quality index with machine learning algorithms [43], the use of ALOS-2 PALSAR-2 and Sentinel-2A imagery to estimate aboveground biomass [44], and the use of synthetic aperture radar (SAR) and light detection and ranging (LiDAR) data to evaluate the flood depth through the application of a normalized difference index [45].…”

Detecting Lithium (Li) Mineralizations from Space: Current Research and Future Perspectives

“…Recently, great attention has been given to non-parametric classifiers such as machine and deep learning algorithms. These techniques have been applied in land-use/land-cover (LULC) problems with great success, outperforming classical algorithms [112][113][114][115]. In 2014, Yu et al [116] conducted a meta-analysis and synthesis of satellite-based land-cover mapping studies, concluding that, from all classification methods (used more than 10 times out of the 1651 analyzed studies), MLAs, namely ensemble classifiers, ANN and SVM, are the methods that achieved a performance better than the average accounted in the 1651 experiments.…”