Study on machine learning classifications based on OLI images

Author

Gao Yan ; Su Fenzhen

Author_Institution

Inst. of Surveying & Mapping, Inf. Eng. Univ., Beijing, China

fYear

2013

fDate

20-22 Dec. 2013

Firstpage

1472

Lastpage

1476

Abstract

Classification for remote sensing images needs to build rules through machine learning. OLI images are useful multi spectral images put into use in 2013. Three kinds of machine learning algorithms were studied for classifying an OLI image in this paper. Samples and 22 features are put in use to test the three kinds of machine learning algorithms. The results are shown as quantitative analysis, visual analysis and feature importance comparison. The results are as follows: In this three machine learning algorithms, using SVM can get the best results, BPNN make the worst results and different classifiers use different features for training and classification.

Keywords

decision trees; geophysical image processing; image classification; learning (artificial intelligence); remote sensing; support vector machines; BPNN; OLI images; SVM; backpropagation neural networks; feature importance comparison; machine learning algorithms; machine learning classification; multispectral images; quantitative analysis; remote sensing image classification; support vector machines; visual analysis; Accuracy; Geometry; Gray-scale; Kernel; Machine learning algorithms; Polynomials; Support vector machines; OLI images; classification; decision tree; machine learning; neural network; support vector machine;

fLanguage

English

Publisher

ieee

Conference_Titel

Mechatronic Sciences, Electric Engineering and Computer (MEC), Proceedings 2013 International Conference on

Conference_Location

Shengyang

Print_ISBN

978-1-4799-2564-3

Type

conf

DOI

10.1109/MEC.2013.6885299

Filename

6885299