Automatic Object Recognition from Satellite Images using Artificial Neural Network

Abstract

Object recognition from satellite images is a very important application for various purposes. Objects can be recognized based on certain features and then applying some algorithm to extract those objects. Basically object recognition is a classification problem. For various remote sensing applications, waterbody acts as an important object which needs to be extracted. It is wise and better if possible, to extract waterbody object automatically from satellite data without any human intervention. This can be achieved using machine learning techniques. Artificial Neural Network (ANN) is such technique which makes machine intelligent by providing learning to it. This intelligent machine can extract objects automatically. This paper presents a methodology to extract waterbody object from satellite data in an automatic manner with the help of ANN. Training and testing dataset have been created by a domain expert which then have been used to train Multi Layer Perceptron (MLP) using Error Back Propagation (EBP) learning algorithm. Confusion matrix and Kappa coefficient have been used for accuracy assessment.

Neural Network Approach for Automatic Landuse Classification of Satellite Images: One-Against-Rest and Multi-Class Classifiers

“…ANN is based on the biological structure of the neurons and their connections in living organisms [2]. It is a massively parallel distributed processor that has a natural propensity for storing experiential knowledge in inter-neuron connection strengths known as synaptic weights by a learning process and making it available for use for prediction [1]. Neurons are arranged in various layers that include an input layer, hidden layers, and an output layer [3].…”

“…MLP is trained using training data before it is applied for prediction [3] and this training is supervised learning where prior information of desired response is known [1]. There are various algorithms proposed for the training of MLP but EBP algorithm is being used widely [4].…”

“…There are various types and variation of ANN classification algorithms such as Multi Layer Perceptron (MLP), Self Organizing Feature Map (SOFM), Radial Basis Functions (RBF), Probabilistic Neural Networks (PNN), Cellular Neural Networks (CNN) etc. MLP is an ANN used very frequently for various applications such as pattern classification, function approximation or prediction and it is universal in the sense that they can approximate any continuous nonlinear function [1]. This paper provides a framework for an automatic classification of landuse using MLP classifier with two different approaches namely one against rest classification (OARC) and multi class classification (MCC) and then compares these two approaches.…”

Automated Identification of River Hydromorphological Features Using UAV High Resolution Aerial Imagery

“…For example, [20,21,22] used ANNs to identify green canopy cover from background soil and shadows. ANNs have also been used successfully to map water bodies [23] and flood extent [24]. However, to the authors’ knowledge the usefulness of ANNs in the classification of river features as part of hydromorphological assessment has not been tested yet.…”

Automatic Extraction of Water Bodies from Landsat Imagery Using Perceptron Model

“…Lippit et al [64] applied MLP to map selective logging sites in deciduous and mixeddeciduous forest in Massachusetts and Li et al [65] and Jensen et al [66] for forest age estimations; Chaudhuri and Parui implemented MLP [67] in defense application to identify target objects. Goswami et al [68] have shown the strength of MLP to identify objects in an image automatically. Their research shows that the perceptron can be effectively used to extract water bodies from satellite data with good accuracy.…”