Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
14 results
Search Results
Item Utility of Landsat Data for Assessing Mangrove Degradation in Muthupet Lagoon, South India(Elsevier, 2018) Subbarayan, S.; Jegankumar, R.; Selvaraj, A.; Jacinth Jennifer, J.; Kulithalai Shiyam Sundar, K.S.S.Mangrove swamps and forests are an essential interface of the coastal zone that provide various ecological and economic services contributing to coastal protection and carbon credits. The ever-changing land use along the coastal tract, especially saltpans and agricultural activities in the mangrove habitats, contribute to the reduction of mangrove swamp sprawl and degrade the mangrove forest. Remote sensing techniques are routinely used to provide spatial-temporal information on mangrove ecosystem distribution, species identification, health status, and population. By adopting supervised classification techniques using the capabilities of indices such as the Normalized Difference Water Index (NDWI) and the Normalized Difference Vegetation Index (NDVI), we attempt to map the spatio-temporal variations of the Muthupet Lagoon regions. The increase of land use changes in the vicinity of the Muthupet Lagoon drastically decrease the freshwater flow and create significant impacts on the mangrove habitat. The study described herein documented degradation of 40.3. ha of dense mangroves from 2013 to 2016, 135.5. ha from 2008 to 2013, and 166. ha from 1999 to 2008 due to high salinity, coastal erosion, and the intrusion of saltpans, aquaculture farmlands, and other human activities. The area of sparse mangroves increased by 38.2. ha between 2013 and 2016, by 42.7. ha from 2008 to 2013, and by 191.3. ha from 1999 to 2008 due to prominent restoration activities. © 2019 Elsevier Inc. All rights reserved.Item Utility of Landsat Data for Assessing Mangrove Degradation in Muthupet Lagoon, South India(Elsevier, 2019) Subbarayan, S.; Jegankumar, R.; Selvaraj, A.; Jennifer, J.; Kulithalai Shiyam Sundar, K.S.S.Mangrove swamps and forests are an essential interface of the coastal zone that provide various ecological and economic services contributing to coastal protection and carbon credits. The ever-changing land use along the coastal tract, especially saltpans and agricultural activities in the mangrove habitats, contribute to the reduction of mangrove swamp sprawl and degrade the mangrove forest. Remote sensing techniques are routinely used to provide spatial-temporal information on mangrove ecosystem distribution, species identification, health status, and population. By adopting supervised classification techniques using the capabilities of indices such as the Normalized Difference Water Index (NDWI) and the Normalized Difference Vegetation Index (NDVI), we attempt to map the spatio-temporal variations of the Muthupet Lagoon regions. The increase of land use changes in the vicinity of the Muthupet Lagoon drastically decrease the freshwater flow and create significant impacts on the mangrove habitat. The study described herein documented degradation of 40.3ha of dense mangroves from 2013 to 2016, 135.5ha from 2008 to 2013, and 166ha from 1999 to 2008 due to high salinity, coastal erosion, and the intrusion of saltpans, aquaculture farmlands, and other human activities. The area of sparse mangroves increased by 38.2ha between 2013 and 2016, by 42.7ha from 2008 to 2013, and by 191.3ha from 1999 to 2008 due to prominent restoration activities. © 2019 Elsevier Inc. All rights reserved.Item Monitoring land use and land cover changes in coastal karnataka(Springer Science and Business Media Deutschland GmbH, 2021) Kumar, M.S.; Venkatesh, V.; Gowthami, S.B.; Anjita, N.A.; Nayana, N.; Regi, L.; Dwarakish, G.S.The dynamics of land use/land cover can be studied by using digital change detection techniques which are highly significant for the evaluation and development of management strategies in a region. The environmental and hydrological processes prevailing in the area can be interpreted only by analyzing the alterations in the past and present land use and land cover classes. In view of this, the present study is executed to analyze the typical land use change in the coastal region over the three decades by analyzing historical and current land use/land cover (LU/LC) datasets. Landsat 5 and Landsat 8 satellite datasets were considered for change detection analysis using unsupervised classification method. K-means algorithm, a widely used unsupervised classification technique, was adopted in this study to classify coastal region of Karnataka for the years 1990 and 2019. The level-II classification was performed on LU/LC raster datasets (Landsat 5 and 8) which segregated the entire study area into ten classes, namely agricultural land, barren land, built-up area, water, forest, fallow or cultivated land, scrub forest, sandy area, swampy forest and wetlands. This study encapsulated that about 40% of the study area was occupied by water body followed by forestry with a percentage of around 30%. Major changes were observed in the barren land and scrub forest between 1990 and 2019, where the barren land was replaced by scrub forest in 2019. The accuracy assessment is performed to analyze the efficiency of the algorithm and the precision of the classified image which showed an accuracy of 81% in 1990 and 84% in 2019 demonstrating the ability of an algorithm to classify reliably. © Springer Nature Singapore Pte Ltd 2021.Item ATGP based Change Detection in Hyperspectral Images(IEEE Computer Society, 2022) Yadav, P.P.; Bobate, N.; Shetty, A.; Raghavendra, B.S.; Narasimhadhan, A.V.Hyperspectral images (HSIs) due to advancements in spatial-spectral resolutions and availability of multi-temporal information is in demand for many remote sensing (RS) applications including change detection (CD). The high dimensionality of HSIs and limited availability of HSI-CD data sets with ground-truth change maps make CD not so easy but a difficult task. Though there are many classical and deep learning (DL) based algorithms to detect changes, the performance of classical algorithms is not up to the satisfactory level and the final performance of DL models depend on efficiency of pre-detection techniques which provide prior knowledge on changed and unchanged areas that are required to get appropriate training samples to learn to detect changes. Classical and DL approaches consider change information at pixel level i.e. pixel to pixel change either by comparing the corresponding pixels alone or with their local neighborhood pixels. Therefore, identification of features for every pixel that relate the most significant information of the whole HSI-CD data in a simple and an efficient way to detect changes effectively is the need of the hour. In addition, there is not much comprehensive study on developing CD algorithms that not only simple to use but also as efficient as that of DL models is available. Therefore, in this paper, an endmember based feature extraction is proposed to detect changes in HSI. An automatic target generation process (ATGP) algorithm is adapted to extract endmembers present in the HSI-CD data set. Then, various spectral matching algorithms are used to measure endmember relations for all the pixels so that dimensionality of the data is reduced as well as the effective features to detect changes can be extracted. The experimental results on three benchmark HSI-CD data sets show that proposed ATGP based change vector analysis (CVA) algorithm yields remarkable results on comparing both with the classical as well as DL based CD approaches. © 2022 IEEE.Item Attention-Based Bitemporal Image Deep Feature-Level Change Detection for High Resolution Imagery(Springer Science and Business Media Deutschland GmbH, 2023) Naik, N.; Chandrasekaran, K.; Venkatesan, M.; Prabhavathy, P.To understand the intricacy of changes on the surface of the land, change detection is an important field in the area of remote sensing. Bitemporal remote sensing images are resourceful information to perform the analysis related to classification and change detection. Most of the architectures proposed for improving the performance of change detection in high resolution images pose a challenge due to composite texture features and finer image details. In this paper, we propose a change detection approach for bitemporal images using supervised learning. Firstly, extraction of the features is performed using a pretrained neural network. Then, the extracted features are provided to a (DSDEN) deep supervised-based difference evaluation network. Then, channel and spatial-based attention components are incorporated for fusing the difference image features with the deep features of raw images for the reconstruction of the final change map. The experimental evaluation on public “LEVIR-CD†dataset demonstrates the effectiveness and superiority over traditional methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.Item DETECTION OF BUILDING INFRASTRUCTURE CHANGES FROM BI-TEMPORAL REMOTE SENSING IMAGES(Institute of Electrical and Electronics Engineers Inc., 2024) Sravya, N.; Kevala, V.D.; Akshaya, P.; Basavaraju, K.S.; Lal, S.; Gupta, D.Change detection (CD) from satellite images is crucial for Earth observation, especially in monitoring urban growth patterns. Recent research has largely focused on using Deep Learning (DL) techniques, particularly variations of Convolutional Neural Network (CNN) architectures. While DL methods have shown promise, many of the models could not preserve the changed areas shape and it fails in predicting the correct edges of changed areas. This paper introduces a CNN based Building Infrastructure Change Detection Network (BICDNet) for predicting changes from bi-temporal remote sensing images. The model leverages a modified Fully Convolutional Siamese-Difference Network to extract detailed features from given images which includes a Multi-Feature Extraction (MFE) block designed to capture features from changed areas of various size within the given input images. To further refine these feature pairs, a modified Atrous Spatial Pyramid Pooling (MASPP) module is integrated, which effectively captures contextual information at multiple scales. The comparison study shows that the proposed BICDNet performs better than the existing CD models. © 2024 IEEE.Item UCDNet: A Deep Learning Model for Urban Change Detection From Bi-Temporal Multispectral Sentinel-2 Satellite Images(Institute of Electrical and Electronics Engineers Inc., 2022) Basavaraju, K.S.; Sravya, N.; Lal, S.; Nalini, J.; Chintala, C.S.; Dell’Acqua, F.Change detection (CD) from satellite images has become an inevitable process in earth observation. Methods for detecting changes in multi-temporal satellite images are very useful tools when characterization and monitoring of urban growth patterns is concerned. Increasing worldwide availability of multispectral images with a high revisit frequency opened up more possibilities in the study of urban CD. Even though there exists several deep learning methods for CD, most of these available methods fail to predict the edges and preserve the shape of the changed area from multispectral images. This article introduces a deep learning model called urban CD network (UCDNet) for urban CD from bi-temporal multispectral Sentinel-2 satellite images. The model is based on an encoder-decoder architecture which uses modified residual connections and the new spatial pyramid pooling (NSPP) block, giving better predictions while preserving the shape of changed areas. The modified residual connections help locate the changes correctly, and the NSPP block can extract multiscale features and will give awareness about global context. UCDNet uses a proposed loss function which is a combination of weighted class categorical cross-entropy (WCCE) and modified Kappa loss. The Onera Satellite Change Detection (OSCD) dataset is used to train, evaluate, and compare the proposed model with the benchmark models. UCDNet gives better results from the reference models used here for comparison. It gives an accuracy of 99.3%, an $F1$ score ( $F1$ ) of 89.21%, a Kappa coefficient (Ka) of 88.85%, and a Jaccard index (JI) of 80.53% on the OSCD dataset. © 1980-2012 IEEE.Item RSCDNet: A Robust Deep Learning Architecture for Change Detection From Bi-Temporal High Resolution Remote Sensing Images(Institute of Electrical and Electronics Engineers Inc., 2023) Deepanshi; Barkur, R.; Suresh, D.; Lal, S.; Chintala, C.S.; Diwakar, P.G.Accurate change detection from high-resolution satellite and aerial images is of great significance in remote sensing for precise comprehension of Land cover (LC) variations. The current methods compromise with the spatial context; hence, they fail to detect and delineate small change areas and are unable to capture the difference between features of the bi-temporal images. This paper proposes Remote Sensing Change Detection Network (RSCDNet) - a robust end-to-end deep learning architecture for pixel-wise change detection from bi-temporal high-resolution remote-sensing (HRRS) images. The proposed RSCDNet model is based on an encoder-decoder framework integrated with the Modified Self-Attention (MSA) andthe Gated Linear Atrous Spatial Pyramid Pooling (GL-ASPP) blocks; both efficient mechanisms to regulate the field-of-view while finding the most suitable trade-off between accurate localization and context assimilation. The paper documents the design and development of the proposed RSCDNet model and compares its qualitative and quantitative results with state-of-the-art HRRS change detection architectures. The above mentioned novelties in the proposed architecture resulted in an F1-score of 98%, 98%, 88%, and 75% on the four publicly available HRRS datasets namely, Staza-Tisadob, Onera, CD-LEVIR, and WHU. In addition to the improvement in the performance metrics, the strategic connections in the proposed GL-ASPP and MSA units significantly reduce the prediction time per image (PTPI) and provide robustness against perturbations. Experimental results yield that the proposed RSCDNet model outperforms the most recent change detection benchmark models on all four HRRS datasets. © 2017 IEEE.Item BCDetNet: a deep learning architecture for building change detection from bi-temporal high resolution satellite images(Springer Science and Business Media Deutschland GmbH, 2023) Basavaraju, K.S.; Hiren, N.S.; Sravya, N.; Lal, S.; Nalini, J.; Chintala, C.S.Change detection is becoming more and more popular technology for the analysis of remote sensing data and is very important for an accurate understanding of changes that are happening in the Earth’s surface. Different Deep Learning methods proposed till now are mainly focused on simple networks which results in poor detection for small changed areas because they can not differentiate between the bi-temporal image’s characteristics. To solve this problem, this article proposes a novel Building Change Detection Network (BCDetNet) for building object change detection and its analysis from bi-temporal high resolution satellite image. The proposed BCDetNet model can detect small change areas with the help of multiple feature extraction block. The proposed BCDetNet model executes building change detection using bi-temporal high resolution satellite images. The proposed BCDetNet model is trained on two publicly available datasets namely LEVIR and WHU change detection(CD) datasets. These datasets contain RGB images with dimensions of (1024 × 1024) and (512 × 512), respectively. The BCDetNet model can learn from scratch during training and performs better than the benchmark change detection models with fewer trainable parameters. The BCDetNet model gives Recall—94.06%, Precision—93.00%, Jaccard score—88.40%, Accuracy—98.73%, F1 score—93.52% and Kappa coefficient—87.05% on LEVIR CD dataset and Recall—89.51%, Precision —92.78%, Jaccard score - 84.38%, Accuracy—96.78%, F1 score—91.06% and Kappa coefficient - 82.12% on WHU CD dataset. This work is a step in the direction of achieving best results in building change detection from high resolution satellite images. © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.Item Spatio-temporal analysis of land use/land cover change detection in small regions using self-supervised lightweight deep learning(Springer Science and Business Media Deutschland GmbH, 2023) Naik, N.; Chandrasekaran, K.; Sundaram, V.; Prabhavathy, P.Change detection (CD) has sparked a lot of scientific interest in recent decades as one of the core concerns in Earth observation. The enhancement of the CD source data with the availability of multitemporal images with varying resolutions provides ample change indicators due to the rapid improvement of satellite sensors. However, precisely detecting real changed locations continues to be a complicated task. CD from remote sensing images (RSI) becomes challenging when the labeled data for supervised learning is unavailable. This article proposes a novel CD framework using a self-supervised learning (SSL) approach to overcome these limitations. First, the superpixel segmentation method of simple linear iterative clustering (SLIC) using a structural similarity index is incorporated to produce a difference image (DI). The change features are extracted to represent the difference information using spatial features between the corresponding superpixels. Second, a parallel clustering algorithm called fuzzy C-means (FCM) separates the DI into three clusters of changed, unchanged, and intermediate classes. The image patches of changed, unchanged and intermediate classes are constructed as training and testing samples. A lightweight deep convolutional neural network (LWDCNN) is trained with the training samples to detect the semantic difference and classify the testing samples into the changed or unchanged class. Finally, merging intermediate and change class labels generates a robust and high-contrast CD map. Numerical experiments were performed on two small regions like the Alappuzha, Kerala, India, and Paris building dataset to demonstrate the usefulness of the proposed approach, achieving an overall accuracy of 98.28% and 96.43% for determining changes effectively. © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.