Research On Algorithm Of Monitoring Points Siting In Mountain Areas

Nowadays, Video monitoring system has been widely used in urban areas. But demand for the video monitoring system that can monitor large zone has not been met. Especially in mountain areas such as in vast mountain areas in the west frontier of China, where big video monitoring systems are needed to help the army to complete border control work more efficiently and easier. Resonable location of montoring points can reduce the cost to build the monitoring system and improve the monitoring area with certain number monitoring points.At present, there two main ways to locate the monitoring points in mountain areas. The first is to site the positions with field measurement. And the second is to find the positions based on DEM and some other geographic data. For comparison, the first way could locate single monitoring point more precisely. But it cannot find a good combnaiton of monitoring point and cost much. And the second way could solve the problem more efficiently and guarantee a combination with much larger coverage area theoretically. Solutions based on SA or GA could solve the problem of siting monitoring points less than 100. However the complexity of the siting problem grow very fast with the increase of the monitoring point number. Siting algorithms base on GA or SA perform worse with the increasing complexity.To complete the work, Digital Elevation Model (DEM) is used as the basic data. Digital Terrain visibility analysis parallel ant colony optimization(ACO) is applied to delevop a new algorithm that could find good combination of positions of monitoring points to reach the max coverage of monitoring scope with the certain amount of monitoring equipment or to get the least number of monitoring equipment under certain monitoring area set efficiently. And the solution of siting monitoring points ACO_VMS and the algorithm is tested in a mountain area about 750km2. So the main work and the result are listed below. i. Visibility analysis algorithm accounting for siting monitoring points is processed Based on traditional visibility algorithm, the characters of video monitoring is parameterized such as the longest distance that the monitoring equipment can distinguish active objects. Compared with the ArcGIS platform, the visibility analysis algorithm presented is more accurate. Then it is applied to simulate the scope of the monitoring equipment. i. Automatically siting monitoring point algorithm based on ACO is developed This paper presents an automatically siting monitoring point algorithm based on ACO combined with the terrain visibility analysis and terrain feature points. And ACO_VMS is tested in the mountain area of 750km2, which needs about 100-200 monitoring points to build a video monitoring system. Compared with traditional siting strategy, ACO_VMS find better position combination with 20-25% larger covered areas by monitoring system. Besides, the paper processed a parallelized ACO_VMS algorithm. Tested with 2-16 compute points, relative speedup are respectively 1.91,3.76,6.94,10.25, and efficiency are respectively 0.96,0.94,0.87,0.64. Obviously parallelized algorithm performs well and increases efficiency notably.i. ACO_VMS is tested in stduy areas with two typical kinds of applicationThe paper presented the algorithm accounting for finding good position combination in mountain areas. The first part is to select candidate points to set monitoring equipment among terrain feature points. Natural conditions such as cliff or steep terrain that is not suitable are also taken into consideration during the selection of candidates points. Second part is to make simulate monitoring equipment by the terrain visibility algorithm presented in i. During the simulation procession, some video monitoring parameters and best solution of the DEM data are optimized. The third part is to find the best position combination after I and ii are completed. Thea mountain area is about 750 km2.500 terrain feature are selected as monitoring point candidates.100-200 monitoring pointed needed to be selected to consist a combination. To evaluate the result of ACO_VMS, the paper process the algorithm base on Simulated Annealing (SA) algorithm to site monitoring points, and also the Regular Grid Random Siting, and the simulation of optimal(SO). Compared with the result of SA, total area covered (TAC) by the result of ACO_VMS is relative 19%-26% higher. TAC of the result of ACO_VMS is higher 20%-25%。ACO_VMS is no less 6.5% than SO.