SpectralSpatial-Aware Unsupervised Change Detection With Stochastic Distances and Support Vector Machines

Autor(es)

Negri, Rogerio Galante ^[1] ; Frery, Alejandro C. ^{[2, 3]} ; Casaca, Wallace ^[4] ; Azevedo, Samara ^{[5, 6]} ; Dias, Mauricio Araujo ; Silva, Erivaldo Antonio ^[7] ; Alcantara, Enner Herenio ^[1]

Afiliação do(s) autor(es)

^[1] Univ Estadual Paulista UNESP, Sci & Technol Inst, Dept Environm Engn, BR-12245000 Sao Jose Dos Campos - Brazil

^[2] Univ Fed Alagoas, Lab Comp Cient & Anal Numer, BR-57072900 Maceio, Alagoas - Brazil

^[3] Xidian Univ, Minist Educ, Key Lab Intelligent Percept & Image Understanding, Xian 710071 - Peoples R China

^[4] Univ Estadual Paulista UNESP, Dept Energy Engn, BR-19274000 Rosana - Brazil

^[5] Univ Fed Itajuba UNIFEI, Dept Nat Resources, BR-35903087 Itajuba - Brazil

^[6] Univ Estadual Paulista UNESP, Dept Math & Comp Sci, Sch Sci & Technol, BR-19060900 Presidente Prudente - Brazil

^[7] Univ Estadual Paulista UNESP, Sch Sci & Technol, Dept Cartog, BR-19060900 Presidente Prudente - Brazil

• Tipo de documento: Artigo Científico
• Fonte: IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING; v. 59, n. 4, p. 2863-2876, APR 2021.
• Citações Web of Science: 1

Resumo

Change detection is a topic of great interest in remote sensing. A good similarity metric to compute the variations among the images is the key to high-quality change detection. However, most existing approaches rely on the fixed threshold values or the user-provided ground truth in order to be effective. The inability to deal with artificial objects such as clouds and shadows is a significant difficulty for many change-detection methods. We propose a new unsupervised change-detection framework to address those critical points. The notion of homogeneous regions is introduced together with a set of geometric operations and statistic-based criteria to characterize and distinguish formally the change and nonchange areas in a pair of remote sensing images. Moreover, a robust and statistically well-posed family of stochastic distances is also proposed, which allows comparing the probability distributions of different regions/objects in the images. These stochastic measures are then used to train a support-vector-machine-based approach in order to detect the change/nonchange areas. Three study cases using the images acquired with different sensors are given in order to compare the proposed method with other well-known unsupervised methods. (AU)

• Processo FAPESP: 18/01033-3 - Investigação e Desenvolvimento de Algoritmos para Detecção de Mudança em Imagens de Sensoriamento Remoto
• Beneficiário: Rogério Galante Negri
• Modalidade de apoio: Auxílio à Pesquisa - Regular
• Processo FAPESP: 13/07375-0 - CeMEAI - Centro de Ciências Matemáticas Aplicadas à Indústria
• Beneficiário: Francisco Louzada Neto
• Modalidade de apoio: Auxílio à Pesquisa - Centros de Pesquisa, Inovação e Difusão - CEPIDs