TY - GEN
T1 - Remote sensing of rough surface parameters using artificial neural network technique
AU - Ishimaru, Akira
AU - Hwang, Jenq Neng
AU - Yoshitomi, Kuniaki
AU - Chen, Jei Shuan
N1 - Funding Information:
This work has been supported by the National Science Foundation and the U.S. Army Research Office. The supercomputer time was granted by the National Science Foundation and the San Diego Supercomputer Center.
Publisher Copyright:
© IEEE 1992.
PY - 1992
Y1 - 1992
N2 - The artificial neural network (ANN) technique is applied to the remote sensing of the rms height and the correlation distance of one-dimensional rough surfaces. The surface is illuminated by a beam wave, and the intensity correlations of the scattered wave at two wavelengths in the specular and backward directions are used to determine the roughness parameters. Scattered intensity correlations calculated by Monte Carlo simulations are used to train the ANN, and two methods, the explicit inversion method and the iterative constrained inversion method, are used to perform the inversion. The technique is applicable to the range of parameters, 0.2 <σ/λ <1.0 and 1.0 < ℓ/λ < 5.0, where σ is the rms height and ℓ is the correlation distance of the surface roughness. An optimum surface area illuminated by the incident beam is approximately 20λ. Both the explicit inverse method and the iterative constrained inversion method give inversion values which are close to the target values. The iterative constrained inversion method appears to give smaller errors, although the required computer time is longer.
AB - The artificial neural network (ANN) technique is applied to the remote sensing of the rms height and the correlation distance of one-dimensional rough surfaces. The surface is illuminated by a beam wave, and the intensity correlations of the scattered wave at two wavelengths in the specular and backward directions are used to determine the roughness parameters. Scattered intensity correlations calculated by Monte Carlo simulations are used to train the ANN, and two methods, the explicit inversion method and the iterative constrained inversion method, are used to perform the inversion. The technique is applicable to the range of parameters, 0.2 <σ/λ <1.0 and 1.0 < ℓ/λ < 5.0, where σ is the rms height and ℓ is the correlation distance of the surface roughness. An optimum surface area illuminated by the incident beam is approximately 20λ. Both the explicit inverse method and the iterative constrained inversion method give inversion values which are close to the target values. The iterative constrained inversion method appears to give smaller errors, although the required computer time is longer.
UR - http://www.scopus.com/inward/record.url?scp=0348155209&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0348155209&partnerID=8YFLogxK
U2 - 10.1109/IGARSS.1992.578344
DO - 10.1109/IGARSS.1992.578344
M3 - Conference contribution
AN - SCOPUS:0348155209
T3 - International Geoscience and Remote Sensing Symposium (IGARSS)
SP - 1072
EP - 1074
BT - IGARSS 1992 - International Geoscience and Remote Sensing Symposium
A2 - Williamson, Ruby
A2 - Stein, Tammy
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 12th Annual International Geoscience and Remote Sensing Symposium, IGARSS 1992
Y2 - 26 May 1992 through 29 May 1992
ER -