iGISc 2019:Papers with Abstracts

Papers
Abstract. Since the network of rainfall gauges and ground radars is generally not dense enough, satellite data have been used to estimate Precipitation (P). These data have the ability to capture the spatial variability pattern of the parameter, but are often inaccurate in relation to the value of the field measured parameter. Therefore, geostatistical methods were evaluated to improve the spatial representativeness of field measurements (FM) and satellite estimates. The work has been made for a hydrological sub region in the Mexican tropic. The geostatistical methods used to interpolate P-FM were ordinary kriging (KO), universal kriging (KU) and regression kriging (RK) as well as the Inverse Distance Weighted (IDW) mechanical interpolator for comparison purposes. Furthermore, the values at the pixel centers of the Tropical Rainfall Monitoring Mission (TRMM) images were interpolated using OK and evaluated using leave-one-out cross validation (LOO-CV). The best LOO-CV evaluated method consisted of the RK interpolation of the point FM taking as auxiliary variable the OK interpolation of the TRMM cell centers. It is concluded that the geostatistical integration between rainfall estimates from satellite data and FM data is promising because satellite information has the ability to capture spatial variability and the point FM add accuracy to the results. These characteristics combined can produce a P product useful for modeling activities and environmental management.
Abstract. S3E2 is a web-based geographic information system (GIS) designed for the visualization and analysis of the socioeconomic segregation of Chile’s elementary education system of Chile. It consists of a frontend developed in JavaScript using ReactJS, React-Redux, Leaflet and D3.js, an API developed in Go and ECHO, and a documentary database man- aged with MongoDB. Data comes from Chile’s Ministry of Education, while the provisions of Law 20.248 serve as indicators of vulnerability. S3E2 graphically shows different segregation indices found in the literature, calculated at the commune level. It also allows visualizing, at this same level, the educational institutions that compose it, their basic information, and time series associated with them. S3E2 is a flexible and fast web-based GIS, with a low cost of implementation, due to the usage of free software -or at least free licensing- tools, thus serving as a template for new web-based GIS in different contexts.
Abstract. Searching clandestine graves is a huge task being conducted by many people around the world. In Mexico, this activity has steadily grown since the disappearance of the 43 students from Ayotzinapa, Gro. leading to the discovery of over a hundred of clandestine graves in the vicinity of Iguala, Gro. In order to facilitate extensive searches, a map of the potential distribution of clandestine graves would be valuable as it can reduce time, cost and effort paid by search brigades. This paper introduces the concept of clandestine space, shows its relation with known grave locations and uses it to map the potential distribution of clandestine graves in Guerrero by means of a machine learning approach.
Abstract. A methodology for the manipulation and analysis of georeferenced points as a result of the theft described in the City of Aguascalientes from 2011 to 2017, the detection of areas with high impact is generated, which end up becoming the inputs for the objective and strategic selection of points of control for timely intervention and constant monitoring of local authorities.
Abstract. Modern Text Mining techniques seek for extract information in useful formats such as georeferences in digital documents. Automatic recognition of location names in texts is usually solved through Named Entity Recognition (NER) systems. Most current NER are based on Machine Learning and have very high accuracy in detection of location entities in digital documents, especially if the texts are in English due to the lack of available an- notated corpora in other languages. However, recent studies are dealing with the challenge of taking the output labels of a NER system and then gather, from a gazetteer, their exact unambiguous geographical coordinates. This is challenging mainly because toponyms use to be very ambiguous, so research in disambiguation methods is relevant. In this paper we describe some of the main ideas towards a method to associate locations with geographical data removing possible confusion between entities with the same name. So far, we have already accomplished Geographic NER and coordinates retrieval but the main research is still in course. We largely discuss about the state of the art around Geoparsing; we explain how our Geographic Entity Recognition module works and finally we describe the research proposal focusing in ambiguity detection.
Abstract. Nowadays, online news sources generate continuous streams of information that includes references to real locations. Linking these locations to coordinates in a map usually requires two steps involving the named entity: extraction and disambiguation. In past years, efforts have been devoted mainly to the first task. Approaches to location disambiguation include knowledge-based, map-based and data-driven methods. In this paper, we present a work in progress for location disambiguation in news documents that uses a vector-semantic representation learned from information sources that include events and geographic descriptions, in order to obtain a ranking for the possible locations. We will describe the proposed method and the results obtained so far, as well as ideas for future work.
Abstract. The analysis of satellite images provides an alternative and complementary method for a better understanding of coral reef ecosystems profitably, with large-scale and near-real- time data. The present study focuses on the presence of coral reef at the Archipelago Espiritu Santo National Park, using high-resolution multispectral images (10 m2) from the Sentinel-2B satellite of the European Space Agency ESA. A Random Forest algorithm was applied to the reflectance bands to estimate bathymetry and classify the seabed in order to assess the coral reef coverage on the island. The results shown are suitable for bathymetry with a variance explained by R2 = 0.895, on the other hand, the classification of bottom type indicates a submerged area of 161.23 ha of coral reef coverage. Reef map- ping, beyond identifying its distribution, has the potential to quantify other parameters that may be important when monitoring these ecosystems.
Abstract. Conservation Area Ejidos de Xochimilco y San Gregorio Atlapulco, which is located in the south of the Mexico Basin, is considered as World Her- itage Site by UNESCO, and a Ramsar Site, that is, it is a Wetland of International Importance. In spite of these international designations, the conservation area has an environmental problem of many years. This article aims to deter- mine the change of vegetation cover and water in the lake area of Xochimilco over a period of 30 years. Remote Sensing facilitates the monitoring of water bodies and vegetation of conservation zones throughout months, even decades. For the period from the year 1987 to year 2016, normalized vegetation indexes NDVI=(NIR-red)/(NIR+Red) and NDWI=(NIR-SWIR)/(NIR+SWIR); and normalized water index MNDWI=(Green-SWIR)/(Green+SWIR), were used together to obtain the changes of areas in vegetation and water in annual images captured by the TM and ETM+ sensors of Landsat 5 and Landsat 7 satellites, respectively. Five classes were obtained (water, water with vegetation, dry vegetation, watered vegetation, and other materials, for instance soil and buildings) with a reliability of 89%.
Abstract. In recent years, the efforts to enhance the analysis of Earth’s surface with satellite imagery have forced the scientific community to develop different techniques and methodologies. The Open Data Cube aims to provide tools to execute multi-temporal analysis and get accurate products, excluding low-quality pixels in small or large areas of study with an accuracy subject to the resolution of the data used for the analysis. This means that we can make use of the full potential of Earth observation data available from satellite data providers, in this document we take a closer look at Landsat Imagery and its applications. The beginning of the implementation of the Open Data Cube platform began in 2018, positioning itself as a valuable source of spatial data for Natural Resources projects in INEGI and seeks to support the decision-making process based on territorial analyzes with great certainty. The use of this technological solution represents a great leap between the traditional visual interpretation of raster data and the automated processing of data in time series.
Abstract. Maya milpa is one of the most important agrifood systems in Mesoamerica, not only because its ancient origin but also due to lead an increase in landscape diversity and to be a relevant source of families food security and food sovereignty. Nowadays, satellite remote sensing data, as the multispectral images of Sentinel-2 platforms, permit us the monitor- ing of different kinds of structures such as water bodies, urban areas, and particularly agricultural fields. Through its multispectral signatures, mono-crop fields or homogeneous vegetation zones like corn fields, barley fields, or other ones, have been successfully detected by using classification techniques with multispectral images. However, Maya milpa is a complex field which is conformed by different kinds of vegetables species and fragments of natural vegetation that in conjunction cannot be considered as a mono-crop field. In this work, we show some preliminary studies on the availability of monitoring this complex system in a region of interest in Yucatan, through a support vector machine (SVM) approach.
Abstract. Photogrammetry encompasses imagery interpretative and measurement method to obtain the shape and location of an object. Since the beginning of the Digital Photogrammetry era (late 80’s), the three-dimensional reconstruction of objects has become one of its fundamental goals [Luhmann et al., 2014]. In the last few years, the professional use of Unmanned Aerial Vehicles (UAVs, also known as drones) focused on this goal has increased and matured relatively fast.
The purpose of this work was to know if and how the UAVs have changed Photogrammetry: how they have modified the Photogrammetric Process and what the necessary bases are to establish a Protocol that provides the necessary steps to make high quality Photogrammetric Surveys of build-ings.
Thanks to the CentroGeo-INEGI project “Desarrollo y Evaluación de Técnicas Avanzadas de Percepción Remota para Alimentar un Catastro Tridimensional”, a team and the author performed a Photogrammetric Survey (of the Ajusco Unit of the National Pedagogical University [UPN] in Mexico City) to evaluate the Protocol.
The author prepared this work as a thesis to obtain the degree of Geomatics Engineer from the National Autonomous University of Mexico (UNAM). For more information, check the TesiUNAM website: shorturl.at/fkBQX
Abstract. Land use change is a global phenomenon that impacts directly to the urban growth and it should be addressed from different disciplines to minimize the potential negative effects of urbanization predicting the spatial urban growth. The urban growth dynamics might be very complicated and difficult to model, nevertheless it is necessary to understand the causes of the dynamics and the dynamics itself to build precise computational models that help to detect problems generated by urban land use change. For that reason, we propose the study of the land use suitability sub-model used by several models to make land use spatial predictions. This sub-model is implemented as a logistic regression based on linear correlations. The problem is that this model is limited to capture a variety of nonlinear relations among variables for prediction and classification purposes. We propose to use an alternative based on Deep Feedforward Networks able to deal with this problem.
In Mexico, the urban growth will increase considerably the number of cities during the next decade where the Mexican population will be concentrated. That means that the generation and study of existing spatio-temporal computational frameworks for studying the Mexican urban growth is very relevant. Therefore we present an initial contribution comparing Deep Feedforward Networks with a Multi-Level Linear Logistic Regression as land suitability models applied to Mexican land use classification. We show that basic deep feedforward models outperform in allocation accuracy to linear logistic regression, and also minimizes the parameters tuned by trial and error.
Abstract. Currently, there are technologies such as Unmanned Aerial Vehicles (UAV) that allow more efficient field work in geology. With the capture of aerial photographs and their processing with algorithms based in Structure from Motion (SfM), it is possible to obtain a position XYZ with their RGB color. The resulting outputs, whether images or point clouds, can be used to recognize patterns of geological layer in stratified walls from color. This semi-automatic segmentation allows to identify subtle changes between lithologies through the transformation of RGB to other color spaces such as CIELab.
The review was done around the issues of color segmentation identified through a search level scheme. These were color information, type of spatial data, segmentation methods and color spaces.
Abstract. Nowadays, remote sensing data taken from artificial satellites require high space com- munications bandwidths as well as high computational processing burdens due to the vertiginous development and specialisation of on-board payloads specifically designed for remote sensing purposes. Nevertheless, these factors become a severe problem when con- sidering nanosatellites, particularly those based in the CubeSat standard, due to the strong limitations that it imposes in volume, power and mass. Thus, the applications of remote sensing in this class of satellites, widely sought due to their affordable cost and easiness of construction and deployment, are very restricted due to their very limited on-board computer power, notwithstanding their Low Earth Orbits (LEO) which make them ideal for Earth’s remote sensing. In this work we present the feasibility of the integration of an NVIDIA GPU of low mass and power as the on-board computer for 1-3U CubeSats. From the remote sensing point of view, we present nine processing-intensive algorithms very commonly used for the processing of remote sensing data which can be executed on-board on this platform. In this sense, we present the performance of these algorithms on the proposed on-board computer with respect with a typical on-board computer for CubeSats (ARM Cortex-A57 MP Core Processor), showing that they have acceleration factors of average of 14.04× ∼14.72× in average. This study sets the precedent to perform satellite on-board high performance computing so to widen the remote sensing capabilities of CubeSats.
Abstract. Fishing is an ancient practice that dates back to at least the beginning of the Upper Paleolithic period about 40,000 years ago. Nowadays, Fishing is one of the most important activities, as it provides a source of food and economic income worldwide. A key challenge in ecology and conservation is to decrease the Illegal, Unreported and Unregulated fishing (IUU). IUU fishing depletes fish stocks, destroys marine habitats, distorts competition, puts honest fishers at an unfair disadvantage, and weakens coastal communities, particularly in developing countries. One strategy to decrease the IUU fishing is monitoring and detecting the fishing vessel behaviors. Satellite–based Automatic Information Systems (S– AIS) are now commonly installed on most ocean–going vessels and have been proposed as a novel tool to explore the movements of fishing fleets in near real time. In this article, we present a dictionary–based method to classify, by using AIS data, between two fishing gear types: trawl and purse seine. The data was obtained from Global Fishing Watch. Our experiments show that our proposal has a good performance in classifying fishing behaviors, which could help to prevent overexploit and improve the strategies of the fisheries management.