Abstract
Abstract. In view of the fact that the deep convolutional neural network has stronger ability of feature learning and feature expression, an exploratory research is done on feature extraction and classification for high resolution remote sensing images. Taking the Google image with 0.3 meter spatial resolution in Ludian area of Yunnan Province as an example, the image segmentation object was taken as the basic unit, and the pre-trained AlexNet deep convolution neural network model was used for feature extraction. And the spectral features, AlexNet features and GLCM texture features are combined with multi-kernel learning and SVM classifier, finally the classification results were compared and analyzed. The results show that the deep convolution neural network can extract more accurate remote sensing image features, and significantly improve the overall accuracy of classification, and provide a reference value for earthquake disaster investigation and remote sensing disaster evaluation.
Highlights
High resolution satellite imagery is an image to obtain information of object detailed on the earth’s surface.High resolution image can be used to classify land use from the process of visual interpretation and classification image(Chen etal.,2015).Visual interpretation and classification image of high resolution image can produce good classification accuracy(Gong etal.,2013;Chen etal.,2007;Chen etal.,2014)
Remote sensing image classification is mainly based on overlay spectral feature classification, researchers have proposed adding texture feature classification, the classification accuracy is improved, but the texture information is still very limited to improve the classification accuracy (Chen etal.,2007;Chen etal.,2014), In recent years, Convolutional Neural Networks (CNN) has made a series of breakthroughs in image classification, target detection, semantic segmentation and face recognition
This paper studies the convolutional neural network for feature extraction and classification of high resolution remote sensing image, taking Ludian post-earthquake Google image with 0.3 meter spatial resolution of images as experimental data, the object of the segmentation image is the basic unit, AlexNet convolution neural network(Krizhevsky etal.,2012) is used for deep feature extraction, which respectively combines with spectral feature and GLCM texture. and the multi- kernel learning is used for fusion feature,SVM classifier is used for classification, The experimental results show that the deep feature can extract more accurate object features and get higher classification accuracy.it shows the shortage of AlexNet in
Summary
High resolution satellite imagery is an image to obtain information of object detailed on the earth’s surface.High resolution image can be used to classify land use from the process of visual interpretation and classification image(Chen etal.,2015).Visual interpretation and classification image of high resolution image can produce good classification accuracy(Gong etal.,2013;Chen etal.,2007;Chen etal.,2014). It is necessary to build classification and recognition related to scene semantics, which are very different from land cover classification.The classification of landsat cover is mainly based on the spectrum, texture, and so on.The main difficulties are the difficulty of the characteristic expression caused by the diversity of the internal spectrum and the problem of the characteristic expression of the mixed spectrum is caused by the mixed pixel.Can the powerful feature extraction capability of the convolution neural network be used to improve the accuracy of surface coverage classification for medium/high resolution remote sensing images? The multi- kernel learning is used for fusion feature,SVM classifier is used for classification, The experimental results show that the deep feature can extract more accurate object features and get higher classification accuracy.it shows the shortage of AlexNet in. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3, 2018 ISPRS TC III Mid-term Symposium “Developments, Technologies and Applications in Remote Sensing”, 7–10 May, Beijing, China the classification of highlighting and lowlighting feature extraction
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
