This paper presents a methodology to detect the coastal erosion and progradation effects in the ‘Atrato River’ delta, located in the Gulf of Urabá in Colombia, using SAR (Synthetic Aperture Radar) images. Erosion is the physical–mechanical loss of the soil that affects its functions and ecosystem services while producing a reduction in its productive capacity. Progradation is the deposition of layers in the basinward direction while moving coastward. Other studies have investigated these two phenomena using optical images, encountering difficulties due to the persistent presence of clouds in this region. In order to avoid the cloud effects, in this study, we used 16 Sentinel 1 SAR images with two different polarizations between 2016 and 2023. First, each image was rescaled from 0 to 255, then the image was despeckled by a deep learning (DL) model. Afterwards, a single RGB image was composed with the filtered polarizations. Next, a classifier with 99% accuracy based on Otsu’s method was used to determine whether each pixel was water or not. Then, the classified image was registered to a reference one using Oriented FAST and Rotated BRIEF (ORB) descriptor. Finally, a multitemporal analysis was performed by comparing every image to the previous one to identify the studied phenomena, calculating areas. Also, all images were integrated to obtain a heatmap that showed the overall changes across eight years (2016–2023) in a single image. The multitemporal analysis performed found that the newly created mouth is the most active area for these processes, coinciding with other studies. In addition, a comparison of these findings with the Oceanic Niño Index (ONI) showed a relative delayed coupling to the erosion process and a coupling of progradation with dry and wet seasons.

Deep Learning methods require immense amounts of labeled data to provide reasonable results. In computer vision applications, and more specifically in despeckling SAR (Synthetic Aperture Radar) images, due to the speckle content, there is no ground truth available. To test the performances of despeckling filters, the common protocol is to synthetically corrupt optical images with a suitable speckle model and then, after filtering, well-known metrics are obtained. Then, filters are tested on actual SAR data. However, even the most elaborated speckle models are far from accounting for the complex mechanisms related to SAR images. In this paper, a methodology to design a realistic dataset is proposed. Actual SAR images of the same scene acquired with the same sensor on different dates, then they are properly co-registered and averaged to get a ground truth-like reference image to objectively evaluate the performance of a despeckling method. To show the benefits of the proposed methodology, a deep learning approach is used to filter the data by using the designed dataset, which will be called the “SAR model”. Then they are compared with the standard protocol by using synthetically corrupted optical images, which will be the “Synthetic model”. One last validation is performed by filtering the same images with FANS, a well-known despeckling filter and compared with the results obtained with autoencoder. The validation on actual SAR data not included in the training phase validates the proposed methodology. From the results shown, it is recommended to test filters on the proposed more realistic dataset.

Speckle reduction in Synthetic Aperture Radar (SAR) images is a crucial challenge for effective image analysis and interpretation in remote sensing applications. This study proposes a novel deep learning-based approach using autoencoder architectures for SAR image despeckling, incorporating analysis of variance (ANOVA) for hyperparameter optimization. The research addresses significant gaps in existing methods, such as the lack of rigorous model evaluation and the absence of systematic optimization techniques for deep learning models in SAR image processing. The methodology involves training 240 autoencoder models on real-world SAR data, with performance metrics evaluated using Mean Squared Error (MSE), Structural Similarity Index (SSIM), Peak Signal-to-Noise Ratio (PSNR), and Equivalent Number of Looks (ENL). By employing Pareto frontier optimization, the study identifies models that effectively balance denoising performance with the preservation of image fidelity. The results demonstrate substantial improvements in speckle reduction and image quality, validating the effectiveness of the proposed approach. This work advances the application of deep learning in SAR image denoising, offering a comprehensive framework for model evaluation and optimization.

When training Artificial Intelligence and Deep Learning models, especially by using Supervised Learning techniques, a labeled dataset is required to have an input with data and its corresponding labeled output data. In the case of images, for classification, segmentation, or other processing tasks, a pair of images is required in the same sense, one image as an input (the noisy image) and the desired (the denoised image) one as an output. For SAR despeckling applications, the common approach is to have a set of optical images that then are corrupted with synthetic noise, since there is no ground truth available. The corrupted image is considered the input and the optical one is the noiseless one (ground truth). In this paper, we provide a dataset based on actual SAR images. The ground truth was obtained from SAR images of Sentinel 1 of the same region in different instants of time and then they were processed and merged into one single image that serves as the output of the dataset. Every SAR image (noisy and ground truth) was split into 1600 images of 512 × 512 pixels, so a total of 3200 images were obtained. The dataset was also split into 3000 for training and 200 for validation, all of them available in four labeled folders.

This article presents a comprehensive dataset combining Synthetic Aperture Radar (SAR) imagery from the Sentinel-1 mission with optical imagery, including RGB and Normalized Difference Vegetation Index (NDVI), from the Sentinel-2 mission. The dataset consists of 8800 images, organized into four folders—SAR_VV, SAR_VH, RGB, and NDVI—each containing 2200 images with dimensions of 512 × 512 pixels. These images were collected from various global locations using random geographic coordinates and strict criteria for cloud cover, snow presence, and water percentage, ensuring high-quality and diverse data. The primary motivation for creating this dataset is to address the limitations of optical sensors, which are often hindered by cloud cover and atmospheric conditions. By integrating SAR data, which is unaffected by these factors, the dataset offers a robust tool for a wide range of applications, including land cover classification, vegetation monitoring, and environmental change detection. The dataset is particularly valuable for training machine learning models that require multimodal inputs, such as translating SAR images to optical imagery or enhancing the quality of noisy data. Additionally, the structure of the dataset and the preprocessing steps applied make it readily usable for various research purposes. The SAR images are processed to Level-1 Ground Range Detected (GRD) format, including radiometric calibration and terrain correction, while the optical images are filtered to ensure minimal cloud interference.

This paper presents an analysis of different despeckling filters applied on both synthetically corrupted optical images and actual Synthetic Aperture Radar (SAR) images. Several authors use optical images as ground truth and then the images are corrupted by using a Gamma model to simulate the speckle, while other approaches use methods like multitemporal fusion to generate a ground truth using actual SAR images, which provides a result somehow equivalent to the one from the common multi look technique. Well-known filters, like local, and non-local and some of them based on artificial intelligence and deep learning, are applied to these two types of images and their performance is assessed by a quantitative analysis. One last validation is performed with a newly proposed method by using ratio images, resulting from the mathematical division (Hadamard division) of filtered and noisy images, to measure how similar the initial and the remaining speckle are by considering its Gamma distribution and divergence measurement. Our findings suggest that despeckling models relying on artificial intelligence exhibit notable efficiency, albeit concurrently displaying inflexibility when applied to particular image types based on the training dataset. Additionally, our experiments underscore the utility of the divergence measurement in ratio images in facilitating both visual inspection and quantitative evaluation of residual speckles within the filtered images.