A Deep Multitask Semisupervised Learning Approach for Chlorophyll-a Retrieval from Remote Sensing Images

This article addresses the scarcity of labeled data in multitemporal remote sensing image analysis, and especially in the context of Chlorophyll-a (Chl-a) estimation for inland water quality assessment. We propose a multitask CNN architecture that can exploit unlabeled satellite imagery and that can be generalized to other multitemporal remote sensing image analysis contexts where the target parameter exhibits seasonal fluctuations. Specifically, Chl-a estimation is set as the main task, and an unlabeled sample’s month classification is set as an auxiliary network task. The proposed approach is validated with multitemporal/spectral Sentinel-2 images of Lake Balik in Turkey using in situ measurements acquired during 2017–2019. We show that harnessing unlabeled data through multitask learning improves water quality estimation performance.

Assessing Vegetation Decline due to Pollution from Waste Cycle Activities by a Multitemporal Remote-Sensing Approach

The work consisted in identifying possible effects from heavy metals (HMs) pollution due to waste disposal activities in three potentially polluted sites located in Basilicata (Italy), where a release of pollutants with values over the thresholds allowed by the Italian legislation was detected. The potential variations in the physiological efficiency of vegetation have been analyzed through the multitemporal processing of satellite images. In detail, Landsat 5 Thematic Mapper (TM) and Landsat 8 Operational Land Imager (OLI) images were used to calculate the Normalized Difference Vegetation Index (NDVI) trend over the years. Then, the multitemporal trends were analyzed using the median of Theil-Sen, a non-parametric estimator particularly suitable for the treatment of remote sensing data, being able to minimize the outlier effects due to exogenous factors. Finally, the subsequent application of the Mann-Kendall test on the trends identified by Theil-Sen slope allowed the evaluation of trends significance and, therefore, the areas characterized by the effects of contamination on vegetation. The application of the procedure to the three survey sites led to the exclusion of the presence of significant effects of HMs contamination on the vegetation surrounding the sites during the years of waste disposal activities.

Indices Extraction from Multitemporal Remote Sensing Data for Mapping Urban Built-Up

Remote sensing data analysis in the cloudy area is still a challenging process. Fortunately, remote sensing technology is fast growing. As a result, multitemporal data could be used to overcome the problem of the cloudy area. Using multitemporal data is a common approach to address the cloud problem. However, most methods only use two data, one as the main data and the other as complementary of the cloudy area. In this paper, a method to harness multitemporal remote sensing data for automatically extracting some indices is proposed. In this method, the process of extracting the indices is done without having to mask the cloud. Those indices could be further used for many applications such as the classification of urban built-up. Landsat-8 data that is acquired during 2019 are stacked, therefore each pixel at the same position creates a list. From each list, indices are extracted. In this study, NDVI, NDBI, and NDWI are used to mapping built-up areas. Furthermore, extracted indices are divided into four categories by their value (maximum, quantile 75, median, and mean). Those indices are then combined into a simple formula to mapping built-up to see which produces better accuracy. The Pleiades as high-resolution remote sensing data is used to assist supervised classification for assessment. In this study, the combination of mean NDBI, maximum NDVI, and mean NDWI result highest Kappa coefficient of 0.771.

Spatiotemporal Change Analysis and Future Scenario of LULC Using the CA-ANN Approach: A Case Study of the Greater Bay Area, China

Land use land cover (LULC) transition analysis is a systematic approach that helps in understanding physical and human involvement in the natural environment and sustainable development. The study of the spatiotemporal shifting pattern of LULC, the simulation of future scenarios and the intensity analysis at the interval, category and transition levels provide a comprehensive prospect to determine current and future development scenarios. In this study, we used multitemporal remote sensing data from 1980–2020 with a 10-year interval, explanatory variables (Digital Elevation Model (DEM), slope, population, GDP, distance from roads, distance from the city center and distance from streams) and an integrated CA-ANN approach within the MOLUSCE plugin of QGIS to model the spatiotemporal change transition potential and future LULC simulation in the Greater Bay Area. The results indicate that physical and socioeconomic driving factors have significant impacts on the landscape patterns. Over the last four decades, the study area experienced rapid urban expansion (4.75% to 14.75%), resulting in the loss of forest (53.49% to 50.57%), cropland (21.85% to 16.04%) and grassland (13.89% to 12.05%). The projected results (2030–2050) also endorse the increasing trend in built-up area, forest, and water at the cost of substantial amounts of cropland and grassland.

ANALYSIS OF THE DEGREE OF GRASSLAND DEGRADATION USING REMOTE SENSING

The objective of this study is to adapt the calculations of the Pasture Degradation Index (GDI) to the Brazilian savannah using medium spatial resolution satellite image for the dry season. Vegetation cover is the main evaluation parameter used to calculate the GDI. The extreme ranges of the grazing class were determined by the NDVI histogram of a single date. Pasture cover was distinguished into five classes called Vegetable Pasture Cover (GVC), derived from NDVI and compared with five other classes derived from field photographs, named Green Coverage Percentage (GCP). The similarity between GVC and GVP demonstrated that GVC can be used to classify pasture cover. As a product of GVC, GDI was obtained. The GDI showed that pasture degradation in Paraíso das Águas is very serious. Extremely severe and Severe degradation occupy 9.28% and 25.22% of the study area, moderate and light degradation of pasture occupy 8.29% and 4.50%, respectively, and the non-degradation area covers 1.43 % of pastures. The results suggest the possibility of applying the GDI, originally developed for natural fields and multitemporal remote sensing data, to evaluate the conditions of the tropical savanna planted fields by means of a unique image.