Model based object recognition through hypothesis and parameter matching

Author

Luo, Fang ; Mulder, N.J.

Author_Institution

ITC, Enschede, Netherlands

fYear

1993

fDate

18-21 Aug 1993

Firstpage

162

Abstract

Presents a model based method to recognize objects. Firstly hypotheses are generated from shape primitives and (their) Boolean operations to predict a complex object, and then are verified by finding the minimum cost in parameter space. A number of optimization techniques are considered and then applied to practical search on real-world data. The authors emphasize parameter estimation and consider the procedure as a numerical optimization problem. A technique for finding global minima is reported, and its efficiency is proven by applying the method for recognition of landuse patches in images of agricultural fields

Keywords

geophysical techniques; geophysics computing; image recognition; image sequences; remote sensing; Boolean operation; agricultural field; complex object; geophysic computing; geophysical measurement technique; global minima; hypothesis; image recognition; land surface; land use remote sensing; landuse; model based method; object recognition; optimization; parameter estimation; parameter matching; parameter space; recognize objects; shape primitive; Cost function; Frequency; Image segmentation; Object recognition; Predictive models; RF signals; Radiometry; Rotation measurement; Shape measurement; Solid modeling;

fLanguage

English

Publisher

ieee

Conference_Titel

Geoscience and Remote Sensing Symposium, 1993. IGARSS '93. Better Understanding of Earth Environment., International

Conference_Location

Tokyo

Print_ISBN

0-7803-1240-6

Type

conf

DOI

10.1109/IGARSS.1993.322514

Filename

322514