P. G. Diwakar

Hierarchical Clustering Algorithm for Land Cover Mapping Using Satellite Images

This paper presents hierarchical clustering algorithms for land cover mapping problem using multi-spectral satellite images. In unsupervised techniques, the automatic generation of number of clusters and its centers for a huge database is not exploited to their full potential. Hence, a hierarchical clustering algorithm that uses splitting and merging techniques is proposed. Initially, the splitting method is used to search for the best possible number of clusters and its centers using Mean Shift Clustering (MSC), Niche Particle Swarm Optimization (NPSO) and Glowworm Swarm Optimization (GSO). Using these clusters and its centers, the merging method is used to group the data points based on a parametric method (k-means algorithm). A performance comparison of the proposed hierarchical clustering algorithms (MSC, NPSO and GSO) is presented using two typical multi-spectral satellite images - Landsat 7 thematic mapper and QuickBird. From the results obtained, we conclude that the proposed GSO based hierarchical clustering algorithm is more accurate and robust.

Integration of speckle de-noising and image segmentation using Synthetic Aperture Radar image for flood extent extraction

Flood is one of the detrimental hydro-meteorological threats to mankind. This compels very efficient flood assessment models. In this paper, we propose remote sensing based flood assessment using Synthetic Aperture Radar (SAR) image because of its imperviousness to unfavourable weather conditions. However, they suffer from the speckle noise. Hence, the processing of SAR image is applied in two stages: speckle removal filters and image segmentation methods for flood mapping. The speckle noise has been reduced with the help of Lee, Frost and Gamma MAP filters. A performance comparison of these speckle removal filters is presented. From the results obtained, we deduce that the Gamma MAP is reliable. The selected Gamma MAP filtered image is segmented using Gray Level Co-occurrence Matrix (GLCM) and Mean Shift Segmentation (MSS). The GLCM is a texture analysis method that separates the image pixels into water and non-water groups based on their spectral feature whereas MSS is a gradient ascent method, here segmentation is carried out using spectral and spatial information. As test case, Kosi river flood is considered in our study. From the segmentation result of both these methods are comprehensively analysed and concluded that the MSS is efficient for flood mapping.

Multi-spectral satellite image classification using Glowworm Swarm Optimization

This paper investigates a new Glowworm Swarm Optimization (GSO) clustering algorithm for hierarchical splitting and merging of automatic multi-spectral satellite image classification (land cover mapping problem). Amongst the multiple benefits and uses of remote sensing, one of the most important has been its use in solving the problem of land cover mapping. Image classification forms the core of the solution to the land cover mapping problem. No single classifier can prove to classify all the basic land cover classes of an urban region in a satisfactory manner. In unsupervised classification methods, the automatic generation of clusters to classify a huge database is not exploited to their full potential. The proposed methodology searches for the best possible number of clusters and its center using Glowworm Swarm Optimization (GSO). Using these clusters, we classify by merging based on parametric method (k-means technique). The performance of the proposed unsupervised classification technique is evaluated for Landsat 7 thematic mapper image. Results are evaluated in terms of the classification efficiency - individual, average and overall.

Predictive modelling of the spatial pattern of past and future forest cover changes in India

This study was carried out to simulate the forest cover changes in India using Land Change Modeler. Classified multi-temporal long-term forest cover data was used to generate the forest covers of 1880 and 2025. The spatial data were overlaid with variables such as the proximity to roads, settlements, water bodies, elevation and slope to determine the relationship between forest cover change and explanatory variables. The predicted forest cover in 1880 indicates an area of 10,42,008 km2, which represents 31.7% of the geographical area of India. About 40% of the forest cover in India was lost during the time interval of 1880–2013. Ownership of majority of forest lands by non-governmental agencies and large scale shifting cultivation are responsible for higher deforestation rates in the Northeastern states. The six states of the Northeast (Assam, Manipur, Meghalaya, Mizoram, Nagaland, Tripura) and one union territory (Andaman & Nicobar Islands) had shown an annual gross rate of deforestation of >0.3 from 2005 to 2013 and has been considered in the present study for the prediction of future forest cover in 2025. The modelling results predicted widespread deforestation in Northeast India and in Andaman & Nicobar Islands and hence is likely to affect the remaining forests significantly before 2025. The multi-layer perceptron neural network has predicted the forest cover for the period of 1880 and 2025 with a Kappa statistic of >0.70. The model predicted a further decrease of 2305 km2 of forest area in the Northeast and Andaman & Nicobar Islands by 2025. The majority of the protected areas are successful in the protection of the forest cover in the Northeast due to management practices, with the exception of Manas, Sonai-Rupai, Nameri and Marat Longri. The predicted forest cover scenario for the year 2025 would provide useful inputs for effective resource management and help in biodiversity conservation and for mitigating climate change.

Development of flood inundation extent libraries over a range of potential flood levels: a practical framework for quick flood response

ABSTRACT The aim of the present study is motivated to build an inundation library for a range of gauge heights, which can be used by decision-makers to anticipate the likely extent of inundation and provide quick response towards warning the habitation at threat. In the present study, two approaches for developing a series of static flood-inundation extent libraries for a range of potential flood levels using historical satellite images, gauge heights and digital elevation model (DEM) are demonstrated. First method is based on the geotagging of gauge height data with corresponding satellite observed inundation extent and the other method supplements the first method in the absence of adequate satellite data-sets by simulating inundation using gauge data and DEM for a range of gauge heights. Simulated inundation extents are validated with the satellite-derived reference inundation extents using spatial statistics, which measure the correspondence between the estimated and observed occurrence of events like probability of detection (POD), false-alarm ratio, and critical success index (CSI). A good correlation between the simulated inundation and satellite-derived inundation extents, with POD varying between 87% and 94%, CSI between 75% and 80% is observed.

Monitoring of fire incidences in vegetation types and Protected Areas of India: Implications on carbon emissions

Carbon emissions released from forest fires have been identified as an environmental issue in the context of global warming. This study provides data on spatial and temporal patterns of fire incidences, burnt area and carbon emissions covering natural vegetation types (forest, scrub and grassland) and Protected Areas of India. The total area affected by fire in the forest, scrub and grasslands have been estimated as 48765.45, 6540.97 and 1821.33 km 2, respectively, in 2014 using Resourcesat-2 AWiFS data. The total CO 2 emissions from fires of these vegetation types in India were estimated to be 98.11 Tg during 2014. The highest emissions were caused by dry deciduous forests, followed by moist deciduous forests. The fire season typically occurs in February, March, April and May in different parts of India. Monthly CO 2 emissions from fires for different vegetation types have been calculated for February, March, April and May and estimated as 2.26, 33.53, 32.15 and 30.17 Tg, respectively. Protected Areas represent 11.46% of the total natural vegetation cover of India. Analysis of fire occurrences over a 10-year period with two types of sensor data, i.e., AWiFS and MODIS, have found fires in 281 (out of 614) Protected Areas of India. About 16.78 Tg of CO 2 emissions were estimated in Protected Areas in 2014. The natural vegetation types of Protected Areas have contributed for burnt area of 17.3% and CO 2 emissions of 17.1% as compared to total natural vegetation burnt area and emissions in India in 2014. 9.4% of the total vegetation in the Protected Areas was burnt in 2014. Our results suggest that Protected Areas have to be considered for strict fire management as an effective strategy for mitigating climate change and biodiversity conservation.

Multi-objective Genetic Algorithm for efficient point matching in multi-sensor satellite image

This paper investigates a new approach for point matching in multi-sensor satellite images. The feature points are matched using multi-objective optimization (angle criterion and distance condition) based on Genetic Algorithm (GA). This optimization process is more efficient as it considers both the angle criterion and distance condition to incorporate multi-objective switching in the fitness function. This optimization process helps in matching three corresponding corner points detected in the reference and sensed image and thereby using the affine transformation, the sensed image is aligned with the reference image. From the results obtained, the performance of the image registration is evaluated and it is concluded that the proposed approach is efficient.

Multi-sensor satellite image analysis using niche genetic algorithm for flood assessment

In this paper, cluster splitting and merging algorithms are used for flood assessment using LISS-III (before flood) and SAR (during flood) images. Bayesian Information Criteria (BIC) is used to determine the optimal number of clusters. Keeping this constraint, the cluster centers are generated using the cluster splitting techniques, namely Mean Shift Clustering (MSC), and Niche Genetic Algorithm (NGA). The merging method is used to group the data points into their respective classes, using the cluster centers obtained from the above techniques. These techniques are applied on the LISS-III and SAR image. Further, the resultant images are overlaid to analyze the extent of the flood in individual land classes. A performance comparison of these techniques (MSC and NGA) is presented. From the results obtained, we deduce that the NGA is efficient.

Multi-Temporal Satellite Image Analysis Using Gene Expression Programming

This paper discusses an approach for river mapping and flood evaluation to aid multi-temporal time series analysis of satellite images utilizing pixel spectral information for image classification and region-based segmentation to extract water covered region. Analysis of Moderate Resolution Imaging Spectroradiometer (MODIS) satellite images is applied in two stages: before flood and during flood. For these images the extraction of water region utilizes spectral information for image classification and spatial information for image segmentation. Multi-temporal MODIS images from “normal” (non-flood) and flood time-periods are processed in two steps. In the first step, image classifiers such as artificial neural networks and gene expression programming to separate the image pixels into water and non-water groups based on their spectral features. The classified image is then segmented using spatial features of the water pixels to remove the misclassified water region. From the results obtained, we evaluate the performance of the method and conclude that the use of image classification and region-based segmentation is an accurate and reliable for the extraction of water-covered region.

Multi-temporal Satellite Image Analysis Using Unsupervised Techniques

This paper presents flood assessment using non-parametric techniques for multi-temporal time series MODIS (Moderate Resolution Imaging Spectro radiometer) satellite images. The unsupervised methods like mean shift algorithm and median cut are used for automatic extraction of water pixel from the image. The extracted results presents a comparative study of unsupervised image segmentation methods. The performance evaluation indices like root mean square error and receiver operating characteristics are used to study algorithm performance. The result reported in this paper provides useful information for multi-temporal time series image analysis which can be used for current and future research.