A Satellite Band Selection Framework for Amazon Forest Deforestation Detection Task
Open Access
- 14 July 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1300-1308
Abstract
The conservation of tropical forests is a topic of significant social and ecological relevance due to their crucial role in the global ecosystem. Unfortunately, deforestation and degradation impact millions of hectares annually, necessitating government or private initiatives for effective forest monitoring. This study introduces a novel framework that employs the Univariate Marginal Distribution Algorithm (UMDA) to select spectral bands from Landsat-8 satellite optical sensor, optimizing the representation of deforested areas. This selection guides a semantic segmentation architecture, DeepLabv3+, enhancing its performance. Experimental results revealed several band compositions that achieved superior balanced accuracy compared to commonly adopted combinations for deforestation detection, utilizing segment classification via a Support Vector Machine (SVM). Moreover, the optimal band compositions identified by the UMDA-based approach improved the performance of the DeepLabv3+ architecture, surpassing state-of-the-art approaches compared in this study. The observation that a few selected bands outperform the total contradicts the data-driven paradigm prevalent in the deep learning field. Therefore, this suggests an exception to the conventional wisdom that 'more is always better'.Keywords
All Related Versions
Funding Information
- the São Paulo Research Foundation (FAPESP) (2014/50937-1, 2015/24485-9, 2017/25908-6, 2018/23908-1, 2019/26702-8, 2023/00811-0, 2023/00782-0)
- National Council for Scientific and Technological Development (CNPq) (465446/2014-0)
- Coordination for the Improvement of Higher Education Personnel (CAPES) (001)
- the National Laboratory for Scientific Computing (LNCC/MCTI, Brazil) for providing HPC resources of the SDumont supercomputer
This publication has 35 references indexed in Scilit:
- Generative Semantic Segmentation2023
- ForestEyes Project: Conception, enhancements, and challengesFuture Generation Computer Systems, 2021
- Spatio-Temporal Deep Learning Approach to Map Deforestation in Amazon RainforestIEEE Geoscience and Remote Sensing Letters, 2020
- Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation2018
- Fully convolutional networks for semantic segmentation2015
- The next Landsat satellite: The Landsat Data Continuity MissionRemote Sensing of Environment, 2012
- SLIC Superpixels Compared to State-of-the-Art Superpixel MethodsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2012
- An introduction and survey of estimation of distribution algorithmsSwarm and Evolutionary Computation, 2011
- Large-Area Classification and Mapping of Forest and Land Cover in the Brazilian Amazon: A Comparative Analysis of ALOS/PALSAR and Landsat Data SourcesIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2010
- ImageNet: A large-scale hierarchical image database2009