TY - GEN
T1 - A new approach to semantic computing with interval matrix decomposition for interpreting deforestation phenomenon
AU - Rachmawan, Irene Erlyn Wina
AU - Kiyoki, Yasushi
N1 - Funding Information:
This work is supported in part by MEXT Grant-in-Aid for the Program for Leading Graduate School, “Global Environmental System Larders (GESL)” and Multimedia Database Laboratory (MDBL), Graduate School of Media and Governance, Keio University. We thank the anonymous reviewers for their valuable comments and suggestions
Publisher Copyright:
© 2019 The authors and IOS Press. All rights reserved.
Copyright:
Copyright 2019 Elsevier B.V., All rights reserved.
PY - 2019
Y1 - 2019
N2 - Deforestation is a major problem in ecosystem degradation and one of the main sources of carbon emission to the atmosphere. The use of multi-temporal satellite remote sensing has proven to be effective means to monitor forest conditions on global scale. The effectiveness of the utilization of remotely sensed images will depend on the analysis model and parameter selection procedure to provide information that meets the requirements of deforestation monitoring. Here we demonstrate the ability of semantic computing for analyzing satellite images that applied to interpreting tropical deforestation. The typical semantic computing works for interpreting numerical or textual data. It still remains challenging for utilizing remote sensing images for environmental monitoring in semantic computing since the data is presented in interval value. Therefore, we proposed interval matrix decomposition for automatically generate semantic projection that address uncertain value in environmental parameters. In this study, independent dimensions of Landsat Thematic Mapper imagery (Landsat-8) and The Phased Array Type L-band SAR-2 (PALSAR-2) were combined to create an integrated interpretation of environmental condition for a study area in the deforestation zone of global tropical forest. We proposed essential semantic interval-dimensions derived from heterogonous satellite images, combination of L-Band SAR and optical, namely: HV gamma-naught, red, green, blue, NIR, SWIR channel; and combinational features: Soil temperature, soil moisture, temporal change density, temporal change velocity, shape and texture. Afterward, semantic computing is employed as analysis model to explain significant knowledge of deforestation activity. The experimental result shows that integrated independent dimensions from both the optical and SAR domains, has potential for presenting for aspect-based deforestation assessment and to enable the design of robust forest monitoring systems.
AB - Deforestation is a major problem in ecosystem degradation and one of the main sources of carbon emission to the atmosphere. The use of multi-temporal satellite remote sensing has proven to be effective means to monitor forest conditions on global scale. The effectiveness of the utilization of remotely sensed images will depend on the analysis model and parameter selection procedure to provide information that meets the requirements of deforestation monitoring. Here we demonstrate the ability of semantic computing for analyzing satellite images that applied to interpreting tropical deforestation. The typical semantic computing works for interpreting numerical or textual data. It still remains challenging for utilizing remote sensing images for environmental monitoring in semantic computing since the data is presented in interval value. Therefore, we proposed interval matrix decomposition for automatically generate semantic projection that address uncertain value in environmental parameters. In this study, independent dimensions of Landsat Thematic Mapper imagery (Landsat-8) and The Phased Array Type L-band SAR-2 (PALSAR-2) were combined to create an integrated interpretation of environmental condition for a study area in the deforestation zone of global tropical forest. We proposed essential semantic interval-dimensions derived from heterogonous satellite images, combination of L-Band SAR and optical, namely: HV gamma-naught, red, green, blue, NIR, SWIR channel; and combinational features: Soil temperature, soil moisture, temporal change density, temporal change velocity, shape and texture. Afterward, semantic computing is employed as analysis model to explain significant knowledge of deforestation activity. The experimental result shows that integrated independent dimensions from both the optical and SAR domains, has potential for presenting for aspect-based deforestation assessment and to enable the design of robust forest monitoring systems.
KW - Computing
KW - Deforestation
KW - Dimensional database
KW - Multispectral image
KW - PALSAR
KW - Retrieval
KW - Semantic
KW - soil degradation
UR - http://www.scopus.com/inward/record.url?scp=85059572953&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85059572953&partnerID=8YFLogxK
U2 - 10.3233/978-1-61499-933-1-353
DO - 10.3233/978-1-61499-933-1-353
M3 - Conference contribution
AN - SCOPUS:85059572953
T3 - Frontiers in Artificial Intelligence and Applications
SP - 353
EP - 368
BT - Information Modelling and Knowledge Bases XXX
A2 - Endrjukaite, Tatiana
A2 - Jaakkola, Hannu
A2 - Dudko, Alexander
A2 - Kiyoki, Yasushi
A2 - Thalheim, Bernhard
A2 - Yoshida, Naofumi
PB - IOS Press
ER -