Comparison of Using SVM and MLP Neural Network for Cloud Detection in MODIS Imagery

Provided by: University of Teesside
Topic: Cloud
Format: PDF
The first step in remotely sensed imagery for all its applications is the detection of the areas which may have been affected by cloud. The approaches for cloud detection can be categorized into classification-based and physical-based. Physical methodologies suffer from some drawbacks as high variability of clouds and dependence of radiance to the emissivity of the surface. In this paper, an innovative solution to the classification-based cloud detection methods has been developed. The selection of bands is based on the physical effect of cloud on both the emissivity and reflection.

Find By Topic