Disasters and accidents always happen somewhere. Therefore, it is necessary to use geodata even in the case of small accident scenarios in order to assess access routes and their impact on the environment. Information support for the safety authorities is necessary in many cases.
Complex traffic accidents require indications of conceivable access routes for the rescue services and, in particular, indications of departure routes for the rescue service in order to be able to supply patients quickly.
Local storms lead to blocked or flooded roads and destroyed infrastructure, which is why rescue forces must be informed about the main areas of operation, access routes and safe areas for the potential accommodation of evacuees.
Large fires - for example in a recycling plant - produce highly contaminated flue gases which, depending on the weather conditions, remain on site (e. g. inversion weather conditions) or drift and diffuse. In addition to the question of the appropriate approaching route for emergency forces, it becomes necessary to warn the population in a short time. For that, weather and meteorological data are of interest.
Large scale disasters also require a lot of spatial data at an early stage in order to be able to identify the already or potentially affected areas and assess the development of the situation.
In many accidents, immediate action is required with very little time planning. The example scenario of a dangerous goods accident illustrates this, but can be easily transferred, for example, to attacks or incidents in production plants with leaking harmful substances. In order to plan the measures at short notice, considerable amounts of data have to be processed in the shortest possible time and "converted" into targeted information for the heads of operations. The basis for this is the investigation carried out by the fire brigades' management cycle, which in such cases are always headed by the fire brigades. In this case, important data must be collected and interpreted for all further steps. Without accurate information, only incomplete assessments of the situation and thus decisions on measures based on incomplete assumptions are possible.
Depending on the dynamics of the situation, the operations manager must quickly assess the dangers for humans, animals, the environment and material assets and decide how to proceed. For example, liquid and gaseous pollutants spread very quickly depending on weather conditions and geographical characteristics of the location. Often the outside temperature determines the formation of flammable atmospheres and reaction products, which drift into the environment in the form of noxious gas clouds - very difficult to predict in the event of changes in wind direction. In addition, some substances are invisible to the human eye and/or odourless or cannot be detected or determined due to the hazard and shut-off limits in accordance with the fire brigade regulations.
As a basis for assessing the situation, considerable amounts of geodata are needed, the processing and provision of which is the core of the subject area. The basis for this is the OGC1-compliant geodata portals of the federal states and municipalities, which provide information on topography, the water network, the environment, population, sewers and direct dischargers, for example. Until recently, however, this data was only updated in multi-annual cycles, which made it necessary to supplement current data sources. Since June 2015, the Sentinel 2 mission, as part of ESA's COPERNICUS system, has been delivering high spatial resolution satellite data in the visible and radar range with 14-day update cycles.
In addition to the geodata, an on-site investigation is necessary to determine the necessary measures. This can be complicated by various factors. For example, substance identification can be problematic in accidents involving dangerous goods transporters from a distance, as the prescribed vehicle signage is not always up-to-date and loading documents are located in the inaccessible driver's cab. This makes the acquisition of the latest data at the place of use indispensable. This includes measurement data obtained via portable sensors (catalytic multi-warning and Ex devices, chemo-sensors or test tubes). In larger locations, terrestrial measuring vehicles (so-called NBC scanners) or the Analytical Task Force (ATF) are used. Such vehicles have photoionization detectors (PID) or optical sensors such as the SIGIS2 system. However, these measurements are often complex and associated with dangers, for example in the case of a potentially 30,000 litre petrol-contaminated vehicle.
New Solution Concepts within the BigGIS Project¶
BigGIS is intended to enable an improved operational environment evaluation by the head of operations by combining the currently acquired mass data with equally extensive and complex archive data. To this end, a new approach of exploration, analysis, integration and visualisation must be developed for the described application scenarios. The goal is a much faster and more reliable solution for providing approach routes to the site, the location of supply areas for emergency personnel moving up, as well as the situation evaluation at the site of operation, including the detection of possible pollutants and the prediction of their temporal and spatial dispersion and the derivation of suitable measures.
In addition to the previously available geo and meteorological data, the BigGIS project will use optical, airborne data to record traffic infrastructure, existing buildings, protected goods and, if necessary, the identification of endangered persons. This data can be captured by a flying robot equipped with multi-spectral or hyperspectral sensors (similar to SIGIS2) and flying autonomously over the scene of the accident. These so-called UAS (Unmanned Aerial Systems) continuously send data over a broad spectrum (visible light up to infrared or far infrared) of the underlying air layers to BigGIS. Hyperspectral sensors often generate data volumes in the gigabyte range for each photo flight and application. BigGIS should be able to handle these data volumes and offer suitable APIs to perform data analysis and integration at scale.
Further images taken with the aid of the flying robots indicate the direction and velocity of propagation of pollutants (liquid, gaseous). This allows for forecasts of the diffusion and drift of the pollutants more accurately than with the usual approaches (Halpaap's club, MET or Memplex club etc.). In principle, it is possible to classify escaping substances or reaction products in pollutant pools or noxious gas clouds by means of absorption spectra in the infrared and far infrared range. Sensors such as Hygas, SIGIS 2 and CHARM are already available as ground stated or aviated systems for that purpose. In the BigGIS project, initial feasibility studies were carried out to test the substance identification of simulated noxious gas clouds with multi-spectral cameras for the visible range, which are suitable for use with small drones due to their relatively low weight.
In order to make efficient use of such drones, the BigGIS project is offers an implementation of autonomous, optimal flight planning by the drone itself.
At the same time, models are being developed that can predict the propagation of a detected gas cloud. The aim is to predict possible future scenarios for the spread of the pollutant, wind directions and strengths (e. g. taken from FeWIS), the geo-positions of neighbouring residential areas and their surrounding structure, for example via satellite images, geographical characteristics of the accident area, depending on the information on the spread of the pollutant, wind direction and strength (e. g. taken from FeWIS). The predicted scenario is to be presented to the operations manager in a compact, understandable way by means of new visual techniques directly with the (unsure) knowledge about the spread of the pollutants and thus enable decision support.