Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography (한국측량학회지)
- Volume 9 Issue 2
- /
- Pages.81-91
- /
- 1991
- /
- 1598-4850(pISSN)
- /
- 2288-260X(eISSN)
Improving Correctness in the Satellite Remote Sensing Data Analysis -Laying Stress on the Application of Bayesian MLC in the Classification Stage-
인공위성 원격탐사 데이타의 분석 정확도 향상에 관한 연구 -분류과정에서의 Bayesian MIC 적용을 중심으로-
Abstract
This thesis aims to improve the analysis accuracy of remotely sensed digital imagery, and the improvement is achieved by considering the weight factors(a priori probabilities) of Bayesian MLC in the classification stage. To be concrete, Bayesian decision theory is studied from remote sensing field of view, and the equations in the n-dimensional form are derived from normal probability density functions. The amount of the misclassified pixels is extracted from probability function data using the thres-holding, and this is a basis of evaluating the classification accuracy. The results indicate that 5.21% of accuracy improvement was carried out. The data used in this study is LANDSAT TM(1985.10.21 ; 116-34), and the study area is within the administrative boundary of Seoul.
본 연구에서는 원격탐사 수치화상 데이타의 분류단계에 가중치를 고려한 Bayesian MLC를 적용하여 그 분석 정확도를 향상시키고자 하였다. 우선, Bayesian 결정법칙을 원격탐사분야 측면에서 분석해 보고 정규확률 밀도함수를 이용하여 n차원으로 확장시켰다. 이 유도과정에서 정의되는 사천확률 항에, 평행육면체 분류결과를 가중치로 적용하여 분류를 실행하였다 그리고 최종적 분류정확도는 확률함수데이타에
Keywords