Securing the borders K.Archanaa Qualification: Under Graduate Geo Informatics Engineering Anna University Chennai-25 ABSTRACT The long diverse Indian border faces threats from an equally diverse group of people ranging from illegal migrants to terrorists, drug smugglers and arms dealers. Many of these people are a direct threat to the country’s internal security. To assist the security personnel the surveillance can be automated to a great extent using unmanned aerial vehicles and integrating the data provided by them at real-time and quantifying change to give out suspected areas of intrusion. INTRODUCTION India is a huge country sharing more than 14000 km of border with her six neighbours. Most of the border areas are inhospitable terrain like the ice-capped mountains of Himalayas or swift rivulets. Infiltration along these borders if they are left inadequately attended becomes easy. Not just migrants, the difficult terrain and inadequate border policing are exploited to the maximum by terrorists, drug traffickers and smugglers as well. The consequences have a telling impact on the country’s security. According to the statistics provided to the Parliament during the period April-November 2005 around 207 terrorists have infiltrated and 1466 terrorist related incidents have taken place in the state of Jammu and Kashmir alone. Needless to say many of these incidents have led to avoidable injuries and fatalities. Therefore the need to improved border policing with the help of latest developments in related technology gains paramount importance. OBJECTIVE The objective of this paper is to put forward a suggestion on how border management can be improved and automated to the maximum possible extent. Surveillance done by various sensors in different platforms is correlated at near real time to identify suspect areas. CURRENT SITUATION The security forces involved in border policing in India still rely on conventional methods to monitor and protect the border like video cameras, floodlights, ground sensors, patrol vehicles, and manned aircraft. These are not very effective always as can be seen from the statistics provided by the premier border managing force of the country the Border Security Force. According to BSF till November 2005, 4668 militants were killed, 11665 were captured and another 1050 of them surrendered. In these operations BSF lost 1349 of its men, while 5429 were injured. v To improve the situation currently efforts are on to procure thermal imagers in large numbers and install them along the border areas. With a range of 4 to 6 km and ability to operate even in total darkness, these are expected to bring down the number of successful infiltrations. However thermal imagers have their own limitations. The thermal imaging system requires a continuous supply of electricity, a certain level of accessibility and the availability of motorable roads. Infiltration mainly occurs along the riverine routes and the hamlets along the char areas. India shares 160 km of problematic riverine border with Pakistan and another 781 km with Bangladesh. Thermal imagers are not viable propositions under such circumstances. India is also making use of remote sensing technology to carry out reconnaissance using various sensors from different platforms including helicopters, airplanes and low orbit remote sensing satellites. However the huge amounts of data collected from these platforms are not used in conjunction with one another because of the obvious difficulties in integrating them together. Therefore we can make use of another emerging technology of Unmanned Aerial Vehicles to exploit developments in remote sensing. Unmanned Aerial Vehicles (UAV s) are remotely piloted or self-piloted aircraft carrying sensors, communication equipment and other payload. They are extensively used in reconnaissance and intelligence gathering roles. UAV s can provide high resolution, precise, near-real-time imagery with their Electro-Optical sensors to the ground control operator. Based on their range-capability they are classified into three categories
UAV s can provide high resolution, precise, near-real-time visible imagery with the help of their electro-optical sensors to the ground control stations. The illegal entrants passing through difficult terrain such as mountains and dense forests can be identified with the help of thermal scanners. These scanners sense the heat energy radiated by the terrain as well as objects. However visible and FLIR (Forward Looking Infra-Red) sensors alone are not sufficient. In inclement weather the microwave sensor Synthetic Aperture Radar (SAR) can be used to accurately map the terrain. This is an active instrument that operates in all weather conditions. METHODOLOGY For border surveillance we need persistent coverage over previously exposed areas. The large amounts of data generated can make sense only if automation is achieved to a certain level in processing it and in identification of possible areas of intrusion. Towards this end we need to equip the UAV s with positioning systems like GPS and feed the data real time into GIS. In Global Positioning System or GPS if the differential corrections are made available to the aircraft via geostationary communication satellites their accuracy becomes much higher. The referenced data can be registered with previously taken imagery and using appropriate change detection algorithms changes can be identified. However there are various other factors that influence the data collected by the sensors that make change detection slightly difficult. Within a given window the atmosphere intervening between a thermal sensor and the ground can increase or decrease the apparent level of radiation coming from the ground. The effect that the atmosphere has on the ground signal will depend on the degree of atmospheric absorption, scatter and emission at the time and place of sensing. Gases and suspended particles in the atmosphere may absorb radiations emitted from ground objects, resulting in a decrease in the energy reaching the thermal sensor. Ground signals can also be attenuated by scattering in the presence of suspended particles. On the other hand gases and suspended particles in the atmosphere may emit radiation of their own, adding to the radiation sensed. Hence, the atmospheric absorption and scattering tend to make the signals from the ground objects seem colder than they are, and atmospheric emission tends to make them warmer than they are. Depending on the atmospheric conditions during imaging, one of these effects will outweigh the other. This will result in a biased sensor output. Both effects are directly related to the atmospheric path length, or distance, through which the radiation is sensed. Meteorological conditions have a strong influence on the form and magnitude of thermal atmospheric effects. Even on a clear day, aerosols can cause major modifications of the signals sensed. Dust, carbon particles, smoke, and water droplets can all modify thermal measurements. These atmospheric constituents vary with site, altitude, time, and local weather conditions. The effects of diurnal temperature variation should also be taken into account. In the case of SAR the changes due to atmospheric and meteorological conditions are restricted to the shorter operating wavelengths of less than 4 cm. The change detection needs to be done only after these effects are taken into consideration. Ideally change detection procedures should involve data acquired by the same or similar sensor and be recorded using the same spatial resolution, viewing geometry, spectral bands, radiometric resolution and the time of the day. One of the more efficient approaches to delineating change in multidate imagery is use of a change-versus-no-change binary mask to guide multidate classification. This method begins with a traditional classification of one image as a reference (time 1). Then, one of the spectral bands from this date is registered to the same band in a second date (time 2). This two-band data set is then analyzed using one of the operations image differencing or ratioing. A threshold is then set to separate areas that have changed between dates from those that have not. This forms the basis for creating a binary mask of change-versus –no-change areas. This mask is then applied to the multiband image acquired at time 2 and only the areas of change are then classified for time 2.A traditional post classification comparison is then performed in the areas known to have changed between dates. A database of recently taken imagery along with all the parameters that affect the signals like the time of the day, temperature, sun angle, weather conditions, etc has to be created. Based on their closeness to the conditions at the time of taking the imagery one set can be used as the base imagery for change detection in the thermal band. The same criteria can be applied for visible light imagery also. For change detection in SAR imagery the most recently taken data can be used. Accurate spatial registration of the various dates of imagery is also a requirement for effective change detection. Registration to within ¼ to ½ pixel is generally required. Clearly, when misregistration is greater than one pixel, numerous errors will result when comparing the images. The amount of change is also quantified and those exceeding a certain threshold value are identified as possible are of suspect. The suspect areas alone are then displayed for verification by the expert who has other details like the position of any authorized personnel in such areas. CONCLUSION Developments in remote sensing, GIS, GPS, etc can thus be fully exploited to safe guard our borders obtaining high resolution multi sensor imagery and integrating at near real time with geographic information system to give that critical edge to the security personnel that can potentially save many peoples lives and limbs. REFERENCES
|