Infinite terrain generation is an important use case for computer graphics, games and simulations. However, current techniques are often procedural which reduces their realism. We introduce a learning-based generative framework for infinite terrain generation along with a novel learning-based approach for level-of-detailing of terrains. Our framework seamlessly integrates with quad-tree-based terrain rendering algorithms. Our approach leverages image completion techniques for infinite generation and progressive super-resolution for terrain enhancement. Notably, we propose a novel quad-tree-based training method for terrain enhancement which enables seamless integration with quad-tree-based rendering algorithms while minimizing the errors along the edges of the enhanced terrain. Comparative evaluations against existing techniques demonstrate our framework’s ability to generate highly realistic terrain with effective level-of-detailing.

Topography plays an important role in the development of infrastructural facilities such as irrigation canals and road networks. To determine the least-cost flow path between two geo-locations, given the grid-based Digital Elevation Models (DEMs) and a unit cost of construction per length, cost of lift to raise water up to a height of 10 meters, set of co-ordinates the resultant flow needs to pass-through. This work develops a generic model namely (i) Highly Scalable Gravitational Flow Model (HSGFM) and (ii) Lift Based Flow Model (LBFM) where anti-gravitational force or pumping is used to lift the water along the surface of the terrain. The algorithm correctness values for the 1KM resolution stand at 82.10%, whereas for 90 meters resolution stands at 82.08%. The average value stands out to be 82.09% proving that both the HSGFM and LBFM algorithms are very effective in practice.

This paper proposes a novel methodology of analyzing mobile Internet of Things (IoT) data by performing spatial and anthropogenic factor-based thematic interactions with it to retrieve interesting patterns that account for the data variation. In order to test out this methodology, a study is conducted by collecting Particulate Matter (PM) data across India using a mobile IoT node, and look into the neighbouring spatial and anthropogenic factors such as human activities, settlement patterns and vegetation profile corresponding to each geo-location of the PM data. By performing the spatial factor analysis on the mobile IoT data, we evaluated the influence of human activities on PM10 levels, most significantly observed for 0<PM10<100, which highlights commercial and industrial zones as primary PM10 contributors. It is also observed that rural settlement regions witness lesser PM10 levels as compared to semi-urban and urban regions. The vegetation profile within 500 m and 1 km buffer zones around all the locations show limited influence over PM10 levels. This study shows the importance of mobile IoT deployment in accounting spatial variation as opposed to traditional stationary IoT deployment for pollution monitoring.

Automated generation and (user) authoring of realistic virtual terrain is most sought for by the multimedia applications like VR models and gaming. The most common representation adopted for terrain is Digital Elevation Model (DEM). In this paper, we propose a novel realistic terrain authoring framework powered by a combination of VAE and generative conditional GAN model. Our framework is an example-based method that attempts to overcome the limitations of existing methods by learning a latent space from a real-world terrain dataset. This latent space allows us to generate multiple variants of terrain from a single input as well as interpolate between terrains while keeping the generated terrains close to real-world data distribution. We also developed an interactive tool that lets the user generate diverse terrains with minimal inputs. We perform a thorough qualitative and quantitative analysis and provide a comparison with other SOTA methods.

Greenhouse gases (GHGs) play an important role in controlling local air pollution as well as climate change. In this study, we retrieved column-averaged dry-air ($X$) mole fractions of carbon dioxide (CO2), methane (CH4), and carbon monoxide (CO) using a ground-based EM27/SUN Fourier transform infrared spectrometer (FTIR). The EM27/SUN spectrometers are widely in use in the COllaborative Carbon Column Observing Network (COCCON). The PROFFAST software provided by COCCON has been used to analyze the measured atmospheric solar absorption spectra. In this letter, the diurnal variation and the time series of daily averaged$X$CO2,$X$CH4, and$X$CO covering the period from December 2020 to May 2021 are analyzed. The maximum values of$X$CO2,$X$CH4, and$X$CO are observed to be 420.57 ppm, 1.93 ppm, and 170.40 ppb, respectively. Less diurnal but clear seasonal changes are observed during the study period.$X$CH4and$X$CO from the Sentinel-5Precursor (S5P)/TROPOspheric Monitoring Instrument (TROPOMI) are compared against the EM27/SUN retrievals. The correlation coefficient for the EM27/SUN retrieved$X$CH4and$X$CO, with the S5P/TROPOMI products, are 0.75 and 0.94, respectively.

Generation of spatial profiles of airborne volcanic ash that is spread at synoptic scales is a problem that directly impacts lives and properties. Robust algorithms are needed to model the distribution using sparse data sampled in the neighborhood of an erupting volcano. Existing Numerical Weather Prediction (NWP) algorithms model the dispersion at coarser spatial resolutions. In this study, we evaluate a geospatial interpolation technique called Kriging [1] to generate prediction and error surfaces. Location and temperature values of ash from 2010 Icelandic eruption were spatially autocorrelated using a stochastic kriging method, known as Empirical Bayesian Kriging (EBK) [7]. The EBK estimates were rigorously validated against NWP for regions with varying sample densities. Subsequently, a method to generate an accurate overlay map using EBK estimates to augment NWP outputs is proposed to aid in categorization and mapping of safety zones.

High resolution Digital Elevation Models(DEMs) are an important requirement for many applications like modelling water flow, landslides, avalanches etc. Yet publicly available DEMs have low resolution for most parts of the world. Despite tremendous success in image super-resolution task using deep learning solutions, there are very few works that have used these powerful systems on DEMs to generate HRDEMs. Motivated from feedback neural networks, we propose a novel neural network architecture that learns to add high frequency details iteratively to low resolution DEM, turning it into a high resolution DEM without compromising its fidelity. Our experiments confirm that without any additional modality such as aerial images(RGB), our network DSRFB achieves RMSEs of 0.59 to 1.27 across 4 different terrains having diverse geographical structures.

Current air pollution monitoring systems are bulky and expensive resulting in a very sparse deployment. In addition, the data from these monitoring stations may not be easily accessible. This paper focuses on studying the dense deployment based air pollution monitoring using IoT enabled low-cost sensor nodes. For this, total nine low-cost IoT nodes monitoring particulate matter (PM), which is one of the most dominant pollutants, are deployed in a small educational campus in Indian city of Hyderabad. Out of these, eight IoT nodes were developed at IIIT-H while one was bought off the shelf. A web based dashboard website is developed to easily monitor the real-time PM values. The data is collected from these nodes for more than five months. Different analyses such as correlation and spatial interpolation are done on the data to understand efficacy of dense deployment in better understanding the spatial variability and time-dependent changes to the local pollution indicators.

In this paper, a novel technique is presented to detect buildings from very high resolution satellite image. This work builds on the learning of ICA based building detection technique from the very high resolution (VHR) multispectral satellite images presented in [1]. The candidate building pixels obtained through ICA are used to extract maximally stable extremal regions (MSER) which are then filtered using geometric properties to obtain final potential buildings. The technique is aimed at reducing false detection at pixel-level and improving object-level performance of [1]. Combining the two works offers an unsupervised building detection technique which is robust towards size, shape, color, types of rooftops and shadows. A wider test image set consisting of 15 images of different dimensions are used to evaluate performance of the complete detection process. The combined technique achieves object-level precision and recall of 80.64% and 83.65% respectively.

Many socio-environmental aspects manifest themselves over space and time, interacting at varying scales of these dimensions. Satellite imagery, available repetitively over a region, provide important clues of these observations across these dimensions. But, also pose enormous challenges in terms of data processing, extracting significant patterns (indicating the underlying processes) and be able to further model them as scientific knowledge of the environmental process. In this paper, an effort has been made to propose a time-variant analysis method based on the shape characteristics of the vegetation response over time to help identify regions of significant changes. The study covers four agricultural-year periods between 2008 and 2012 over the district of West Godavari, in south of India. This approach shows that the effect of 2009 drought year on the agricultural practices vary spatially depending on the access to resources and the time-lag that manifests itself in such processes. In this study, we also find that nearly 80% of the region is well endowed and hence resilient to the climatic vagaries.

Data fusion is a prevalent method to extract the best combination of satellite images from different modalities - spectral, spatial, temporal. A new method of object-based fusion of high resolution multispectral (MS) and panchromatic (PAN) images is proposed in this paper, which emphasizes on spectral characteristics preservation. It is a hybrid approach where individual objects detected in the images are considered for mapping data and information transfer is done on a per-pixel basis. In addition, the paper proposes a quantitative assessment measure to assess the spectral distortion of the fused outcomes. The quantitative results demonstrate that the proposed method is better in terms of preserving spectral characteristics as compared to other widely used methods such as Principle Component (PC) based fusion, Intensity Hue Saturation (IHS) based fusion and Color Normalization (CN) or Brovey.

In the recent past there is increased interest in detection and extraction of buildings using object based approaches on given high to very high resolution imagery. In this paper, we introduce a new unsupervised approach to detect buildings from the very high resolution (VHR) multispectral satellite image. Independent component analysis (ICA) followed by Otsu thresholding is used for extraction of multicoloured buildings of diversified size and shape. QuickBird image of Legaspi city has been used to analyze the technique and detection results obtained from three subset images indicate average detection percentage of 84.35 % along with 38.19 % branch factor value. Object-level evaluation results give 72.48% for fully detected buildings count.

The traditional pixel-based classification totally relies on spectral information and neglects the spatial information. These methods when applied on very high-resolution imagery get confused because of the increased variability implicit within the data and thus leads to lower classification accuracies. The object-based image analysis (OBIA) is advantageous to deal with objects that are composed of homogeneous pixels. This paper aims at automatically extracting buildings from very high-resolution satellite imagery using Object Based Image Analysis(OBIA). The algorithm uses an active contour model called chan-vese segmentation to create objects from the image. Objects representing vegetation or trees are removed by subtracting NDVI mask from the segmented output. The detected objects are further filtered based on regional properties like minimum area, width of object etc. The results are promising with 74–77% of the buildings getting detected as objects.

We have addressed the problem of visualization of vector building data on top of a terrain. It tries to make it more realistic by automated draping of building textures from geo-tagged images which are captured from a cell phone camera with a built-in GPS. We use the properties of the images to tag them to the corresponding 2D polygon footprint by using the camera pose, and the cameras position to automate our process. The elevation data captured from NASAs SRTM project, GTOPO, etc can be used to render 3D terrains using existing GPU based Level of detail (LOD) algorithms[1] [2] [3]. The vector data of buildings along with their height attributes can be obtained from various Lidar based algorithms or from DEM data[4] [5]. A scene graph is used as the data structure which is used to render these vector based graphics models, supported by osgEarth [6].

Image Classification techniques are aimed at improving the class accuracies which are affected by the occurrence of mixed pixels in the remotely sensed data. Improving the labeling accuracies for the mixed pixels regions will increase the global class accuracies. Spectral unmixing has been used to decompose the mixed pixel regions into its constituent endmembers, and a corresponding fractional abundance for each endmember. The unmixing approaches are approximated based on spectral behavior, but ignore the spatial neighborhood. The data values at the pixel along with its spatial neighborhood are good indicators of the image characteristics including atmospheric conditions and need to be considered. In the current research, we propose a spatio-spectral framework that improves the classification accuracy and demonstrates its utility by improving the labels of the detected mixed regions for MODIS data and validated with AWIFS (APLULC 2005) derived land cover dataset.

Dynamic reconfiguration of mobile networks is gaining importance in context of new generation networks and has been independently analyzed from the perspectives of throughput optimization, energy conservation and load balancing. We propose a method that performs dynamic reconfiguration of transmission powers of eNodeBs based on information from terminal measurements for optimization of throughput and better management of energy which are dependent on the load conditions in the network. The possibility of powering off of eNodeBs in low/moderate load conditions is investigated and simulation results are presented. Simulation results show improvement in signal quality and throughput while reducing the energy consumption by around 20% in low load conditions. Comparison of the results of minimizing two different cost functions with the same bi-objective but different priorities demonstrates the importance of considering coverage (over network space) and throughput together during reconfiguration.

Seasonal characteristics and crop growth information is of great utility for crop management. The primary occupation in India being agriculture, it is important to devise quick and reliable methods that will help in making decisions affecting agricultural practices at a macro level more efficiently. In this paper, the authors have devised an algorithm to derive a seasonal calendar from the time series data of a moderate resolution satellite, MODIS which is one step short of producing a crop calendar. The method proposed involves filtering time series data using Local Maximum Fitting, finding maximum and minimum points on the time series profile, calculating phenological parameters namely start of season, mid of season, end of season and seasonal amplitude and finally clustering these phenological parameters to obtain a cluster center which is representative of a particular season/cropping practice. The initial results are promising as they are similar to the information available in the form of handouts of the regions under study. A full scale validation involving field visits and comparison with statistical data from government sources will prove the utility of this product. Nevertheless, this work demonstrates the utility of time series MODIS data for obtaining phenological parameters.

Automated road network extraction from remotely sensed imagery is of importance in the context of road databases creation, refinement and updating. Substantial amount of research has been carried out to extract road network from satellite imagery in the photogrammetric and computer vision communities. However, little research has been conducted on utility of multi-temporal satellite images in the context of road extraction. This paper first proposes a simple scheme to detect vehicles using high resolution multi-temporal images of a geographic location and uses the detected road seeds in Fourier descriptor based road tracking algorithm to extract the road network. Performance of the proposed method on CARTOSAT-2 images is discussed.