Land Cover/Land Use Mapping of LISS IV Imagery Using Object-Based Convolutional Neural Network with Deep Features

S Rajesh,T Gladima Nisia,S Arivazhagan,R Abisekaraj

doi:10.1007/s12524-019-01064-9

S Rajesh, T Gladima Nisia + Show 2 more

Open Access

https://doi.org/10.1007/s12524-019-01064-9

Copy DOI

Abstract

The paper proposes a new method for classifying the LISS IV satellite images using deep learning method. Deep learning method is to automatically extract many features without any human intervention. The classification accuracy through deep learning is still improved by including object-based segmentation. The object-based deep feature learning method using CNN is used to accurately classify the remotely sensed images. The method is designed with the technique of extracting the deep features and using it for object-based classification. The proposed system extracts deep features using pre-defined filter values, thus increasing the overall performance of the process compared to randomly initialized filter values. The object-based classification method can preserve edge information in complex satellite images. To improve the classification accuracy and to reduce complexity, object-based deep learning technique is used. The proposed object-based deep learning approach is used to drastically increase the classification accuracy. Here, the remotely sensed images were used to classify the urban areas of Ahmadabad and Madurai cities. Experimental results show a better performance with the object-based classification.

Highlights

Classifying different areas of remote sensing image has a wide variety of applications in fields such as land cover mapping and detection, water resource detection, agricultural usage, wetland mapping, geological information and urban and regional planning
In the remote sensing field, several researches are done for image classification using deep learning models, such as stacked autoencoder (SAE) and convolutional neural network (CNN)
The challenge was efficiently handled by choosing CNN for automatic feature learning with the help of pre-defined filter values

Summary

Introduction

Classifying different areas of remote sensing image has a wide variety of applications in fields such as land cover mapping and detection, water resource detection, agricultural usage, wetland mapping, geological information and urban and regional planning. In the remote sensing field, several researches are done for image classification using deep learning models, such as stacked autoencoder (SAE) and convolutional neural network (CNN). The CNN algorithm is popular for high-resolution image classification due to its effectiveness in spatial feature exploration (Zhao and Du 2016a, b; Zhao et al 2015; Yue et al 2015; Chen et al 2014a). Since the methodology results in many features, some of which are found to be not useful, the best among the wavelet packet statistical and wavelet packet co-occurrence textural feature sets is selected using genetic algorithm (Rajesh et al 2013). Our proposed system replaces the so-far carried-out works for classification of LISS IV image with a deep learning approach. In order to still increase the efficiency, the obtained deep features are combined with the object-based textural features. The scene details of the area are as follows: Satellite/sensor: Resolution: Band 2 (green): Band 3 (red): Band 4 (near-infrared): IRS P6/LISS IV

Proposed Method

Experimental Study and Results

Findings

Conclusion