Cloud Masking Technique for High-Resolution Satellite Data: An Artificial Neural Network Classifier Using Spectral &
- PDF / 2,383,056 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 101 Downloads / 187 Views
RESEARCH ARTICLE
Cloud Masking Technique for High-Resolution Satellite Data: An Artificial Neural Network Classifier Using Spectral & Textural Context Ramakrishna M. V. Malladi1 • Adiba Nizami2 • Manju S. Mahakali2 • Barla Gopala Krishna3 Received: 12 September 2018 / Accepted: 9 October 2018 Ó Indian Society of Remote Sensing 2018
Abstract Cloud masking is a very important application in remote sensing and an essential pre-processing step for any information derivation applications. It helps in estimation of usable portion of the images. Many popular spectral classification techniques rely upon the presence of a short-wave infrared band or bands of even higher wavelength to differentiate between clouds and other land covers. However, these methods are limited to sensors equipped with higher wavelength bands. In this paper, a generic and efficient technique is attempted using the Cartosat-2 series (C2S) satellite which is having high-resolution multispectral sensor in the visible and near-infrared bands. The methodology is based on textural features from the available spectral context, and using a feedforward neural network for the classification is proposed. The method was shown to have an overall accuracy of 97.98% for a large manually pre-classified validation dataset with more than 2 million data points. Experimental results and cloud masks generated for various scenes show that the method may be viable as a reasonable cloud masking algorithm for C2S data. Keywords Cloud masking High-resolution satellite data Artificial neural network Feed forward network GLCM texture Image classification
Introduction Remote sensing satellite imagery has been extensively used to monitor many physical phenomena on earth. Typically, more than 50% of the earth’s surface is covered by clouds (King et al. 2013) and a thick cloud can block almost all information from the surface. Applications monitoring climate change, flood detection, land cover classification (Li et al. 2012), change detection, solar radiation intensity (Matuszko 2012), ocean colour data processing (Chen and Barla Gopala Krishna: Deputy Director, DPPA & WAA (Retd.), NRSC, Hyderabad. & Manju S. Mahakali [email protected] 1
Mechanical Engineering Department, School of Technology, Pandit Deendayal Petroleum University, Raisan, Gandhinagar, Gujarat 382007, India
2
National Remote Sensing Centre (NRSC), Indian Space Research Organization (ISRO), Hyderabad, India
3
NRSC, Hyderabad, India
Zhang 2015), etc., suffer from imperfect masking of clouds. Accurate detection of clouds in scenes is therefore desired, and many automated algorithms based on spectral, spatial or temporal contexts have been proposed over the years. Automatic cloud cover assessment (ACCA; Irish 2000), a threshold-based method using the spectral and thermal bands of Landsat 7 Enhanced Thematic Mapper Plus (ETM ?) and Fmask (Function of mask; Zhu and Woodcock 2012), an object-based method proposed for Landsat 7 imagery provided reasonably high accuracies, both over 90% but still
Data Loading...
