Evaluation of Semantic Segmentation Methods for Land Use with Spectral Imaging Using Sentinel-2 and PNOA Imagery

Pedrayes, Oscar D.; Lema, Darío G.; García, Daniel; Usamentiaga, Rubén; Alonso, Ángela

doi:10.3390/rs13122292

Cited by 15 publications

(5 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ASPP module first appeared in the semantic segmentation algorithm DeepLab V2, consisting of 3 × 3 convolution of four different expansion coefficients. Subsequently, ASPP was applied to many image classifications tasks as an independent module ( Yuan et al., 2019 ; Liu et al., 2021 ; Pedrayes et al., 2021 ). The branch structure inside ASPP is not invariable, and designers often adopt different parameter configurations according to different application scenarios.…”

Section: Multi-scale Feature Extraction Methods Based On Deep Convolu...mentioning

confidence: 99%

Crop classification in high-resolution remote sensing images based on multi-scale feature fusion semantic segmentation model

Gao

Wang

2023

Front. Plant Sci.

View full text Add to dashboard Cite

The great success of deep learning in the field of computer vision provides a development opportunity for intelligent information extraction of remote sensing images. In the field of agriculture, a large number of deep convolutional neural networks have been applied to crop spatial distribution recognition. In this paper, crop mapping is defined as a semantic segmentation problem, and a multi-scale feature fusion semantic segmentation model MSSNet is proposed for crop recognition, aiming at the key problem that multi-scale neural networks can learn multiple features under different sensitivity fields to improve classification accuracy and fine-grained image classification. Firstly, the network uses multi-branch asymmetric convolution and dilated convolution. Each branch concatenates conventional convolution with convolution nuclei of different sizes with dilated convolution with different expansion coefficients. Then, the features extracted from each branch are spliced to achieve multi-scale feature fusion. Finally, a skip connection is used to combine low-level features from the shallow network with abstract features from the deep network to further enrich the semantic information. In the experiment of crop classification using Sentinel-2 remote sensing image, it was found that the method made full use of spectral and spatial characteristics of crop, achieved good recognition effect. The output crop classification mapping was better in plot segmentation and edge characterization of ground objects. This study can provide a good reference for high-precision crop mapping and field plot extraction, and at the same time, avoid excessive data acquisition and processing.

show abstract

Section: Multi-scale Feature Extraction Methods Based On Deep Convolu...mentioning

confidence: 99%

Crop classification in high-resolution remote sensing images based on multi-scale feature fusion semantic segmentation model

Gao

Wang

2023

Front. Plant Sci.

View full text Add to dashboard Cite

show abstract

“…The proposed model is based on the LeNet architecture 27 , which is a very simple CNN with few convolutional layers, causing the computational cost associated with training this CNN to be very low. Through several experimental tests, a balance was reached between the complexity of the architecture and the robustness of the model, achieving results that outperformed larger and more complex architectures in similar tasks 6,14,15,26,28 .…”

Section: Proposed Model Setup and Classificationmentioning

confidence: 99%

“…In this research they showed that the best results were achieved by a deep neural network containing two-dimensional convolutional layers (Conv2D) and principal component analysis (PCA). A common limitation of the above work, as others in the literature 6,14,15 , is the high computational resources needed, especially when dealing with large and complex datasets, this due to complex CNNs architectures 6,14,15 . High computational resources are not always available for all research centers, making the task of LULC classification harder, if not impossible.…”

Section: Introductionmentioning

confidence: 99%

“…To deal with the different issues highlighted above, in this research we propose a simple but robust CNN architecture, capable of mapping 11 LULC in Andean and mountainous areas. We use a simple architecture that overcomes several challenges associated with more complex CNN architectures, such as high hardware costs 15 and longer wait times for results. By using this approach, we aim to provide an efficient and accessible solution for LULC classification using DL techniques 16 .…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Land cover classification of Andean sub-basins in Colombia based on Sentinel-2 satellite images and deep learning

Arrechea-Castillo,

Solano-Correa,

Muñoz-Ordóñez

et al. 2023

Geospatial Informatics XIII

View full text Add to dashboard Cite

The Las Piedras River sub-basin, located in the department of Cauca, Colombia, is very important for the region, especially for the capital (Popayán). This is because this sub-basin contributes around 68.17% of the water supply for the city. To guarantee continuity of this resource, good management of the Water Ecosystem Services (WES) must be carried out. To this aim, periodic environmental assessments of the water resource in the region are necessary. Such Environmental Assessment WES (EAWES) is possible when an accurate and up-to-date land cover map is available. However, obtaining such a product is quite complex due to the heterogeneous conditions both in the land cover and orography of the studied region. Another impacting factor is the weather conditions of the region, that make it difficult to access the areas and/or to acquire information for land cover mapping. This research proposes a robust model, based on deep learning and Sentinel2 satellite images, able to perform a land cover classification with reliable accuracy (>90%) at a low computational cost. A variant of a LeNet convolutional neural network has been used together with features extracted from original spectral bands, radiometric indices and a digital elevation map. Preliminary results show an overall accuracy of 95.49% in the training data and 96.51% in the validation one.

show abstract

“…Unfortunately, the excellent resolution achieved in the visible and near-infrared (VIS-NIR) spectral bands by Sentinel-2 may not be enough in some applications, especially to make use of the information given by the LR and VLR bands located in the infrared (IR) and Short-wave IR (SWIR) spectrum. These bands are suitable for a wide range of applications, such as environmental studies [9,10] and the production of land cover maps [11,12] which motivates the study of algorithms to enhance the LR bands' spatial resolution.…”

Section: Introductionmentioning

confidence: 99%

Single-Image Super-Resolution of Sentinel-2 Low Resolution Bands with Residual Dense Convolutional Neural Networks

2021

View full text Add to dashboard Cite

Sentinel-2 satellites have become one of the main resources for Earth observation images because they are free of charge, have a great spatial coverage and high temporal revisit. Sentinel-2 senses the same location providing different spatial resolutions as well as generating a multi-spectral image with 13 bands of 10, 20, and 60 m/pixel. In this work, we propose a single-image super-resolution model based on convolutional neural networks that enhances the low-resolution bands (20 m and 60 m) to reach the maximal resolution sensed (10 m) at the same time, whereas other approaches provide two independent models for each group of LR bands. Our proposed model, named Sen2-RDSR, is made up of Residual in Residual blocks that produce two final outputs at maximal resolution, one for 20 m/pixel bands and the other for 60 m/pixel bands. The training is done in two stages, first focusing on 20 m bands and then on the 60 m bands. Experimental results using six quality metrics (RMSE, SRE, SAM, PSNR, SSIM, ERGAS) show that our model has superior performance compared to other state-of-the-art approaches, and it is very effective and suitable as a preliminary step for land and coastal applications, as studies involving pixel-based classification for Land-Use-Land-Cover or the generation of vegetation indices.

show abstract

Evaluation of Semantic Segmentation Methods for Land Use with Spectral Imaging Using Sentinel-2 and PNOA Imagery

Cited by 15 publications

References 28 publications

Crop classification in high-resolution remote sensing images based on multi-scale feature fusion semantic segmentation model

Crop classification in high-resolution remote sensing images based on multi-scale feature fusion semantic segmentation model

Land cover classification of Andean sub-basins in Colombia based on Sentinel-2 satellite images and deep learning

Single-Image Super-Resolution of Sentinel-2 Low Resolution Bands with Residual Dense Convolutional Neural Networks

Contact Info

Product

Resources

About