TY - CHAP
T1 - Computational intelligence and its application in remote sensing
AU - Ressom, Habtom
AU - Miller, Richard L.
AU - Natarajan, Padma
AU - Slade, Wayne H.
N1 - Publisher Copyright:
© 2005 Springer. Printed in the Netherlands.
PY - 2005
Y1 - 2005
N2 - This chapter introduces to the fundamentals of computational intelligence and its applications in remote sensing of coastal aquatic environments. It gives insight into the use of popular computational intelligence paradigms in estimating coastal bio-optical parameters and phytoplankton primary production. Several advantages give neural networks an edge over other empirical models as candidates for the next generation of ocean color algorithms. Particular advantages are: (i) resilience to noise; (ii) ease of construction without requiring detailed knowledge of the underlying relationships between remotely sensed data and the desired biogeochemical/bio-optical parameters; (iii) the flexibility to incorporate difficult to manage variables; (iv) efficiency in dealing with non-linearity; and, (v) robustness with respect to redundant inputs. Neural networks can also efficiently cope with large amounts of data and lend themselves to process complex datasets, where different information sources are combined. As large volume and high dimensional data are being generated by the rapidly expanding remote sensing technology, the number of reported applications of computational intelligence based techniques is steadily increasing. Some potential applications of computational intelligence in remote sensing of aquatic coastal environments could include monitoring dissolved oxygen, harmful algal blooms, coral reefs, and submerged aquatic vegetation. With the increasing demand, however, comes the need for further improvements that can make CI based implementation in remote sensing more efficient. Key improvements include: (i) enhanced computational power to handle the high dimensionality and large volume data; (ii) higher temporal, spectral, and spatial resolution and access to near real time data; (iii) increased availability of in situ bio-optical measurements coincident with remotely sensed data (note that like other empirical models, CI-based models are only as good as the dataset to which they are applied; hence, the quality of the data collected is very important); and, (iv) advances in computational intelligence techniques to enhance their speed and make them more accessible to the user.
AB - This chapter introduces to the fundamentals of computational intelligence and its applications in remote sensing of coastal aquatic environments. It gives insight into the use of popular computational intelligence paradigms in estimating coastal bio-optical parameters and phytoplankton primary production. Several advantages give neural networks an edge over other empirical models as candidates for the next generation of ocean color algorithms. Particular advantages are: (i) resilience to noise; (ii) ease of construction without requiring detailed knowledge of the underlying relationships between remotely sensed data and the desired biogeochemical/bio-optical parameters; (iii) the flexibility to incorporate difficult to manage variables; (iv) efficiency in dealing with non-linearity; and, (v) robustness with respect to redundant inputs. Neural networks can also efficiently cope with large amounts of data and lend themselves to process complex datasets, where different information sources are combined. As large volume and high dimensional data are being generated by the rapidly expanding remote sensing technology, the number of reported applications of computational intelligence based techniques is steadily increasing. Some potential applications of computational intelligence in remote sensing of aquatic coastal environments could include monitoring dissolved oxygen, harmful algal blooms, coral reefs, and submerged aquatic vegetation. With the increasing demand, however, comes the need for further improvements that can make CI based implementation in remote sensing more efficient. Key improvements include: (i) enhanced computational power to handle the high dimensionality and large volume data; (ii) higher temporal, spectral, and spatial resolution and access to near real time data; (iii) increased availability of in situ bio-optical measurements coincident with remotely sensed data (note that like other empirical models, CI-based models are only as good as the dataset to which they are applied; hence, the quality of the data collected is very important); and, (iv) advances in computational intelligence techniques to enhance their speed and make them more accessible to the user.
UR - http://www.scopus.com/inward/record.url?scp=84979788546&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84979788546&partnerID=8YFLogxK
M3 - Chapter
AN - SCOPUS:84979788546
T3 - Remote Sensing and Digital Image Processing
SP - 205
EP - 227
BT - Remote Sensing and Digital Image Processing
PB - Springer International Publishing
ER -