Adaptive Algorithm of Moving Objects Detection in Video Monitoring

: pp. 168 - 172
Lviv Polytechnic National University
Lviv Polytechnic National University

The paper based on the analysis of existing methods of detecting moving objects offer combined detection algorithm that adapts to such destabilizing factors as: the noise of the equipment (photoelectron matrix converter, amplifier, analog-to-digital converter), change lighting scenes, change the background. The essence of the proposed algorithm is used two parallel detection methods: the method of background subtraction and frame difference method, which allows the algorithm to work effectively in cases of dynamic background. Statistical evaluation of parameters of the model background, which is performed in the process of learning algorithm for each pixel of the video frame in order to determine the optimal threshold binarization algorithm enables this without setting its parameters with just about any camera, regardless of the level of intrinsic noise eyo, which each camera is different. To increase the efficiency of detection of moving objects in the final stage of the algorithm is filtered binary mask by applying morphological operations to remove wrongly marked pixels and combining pixel marked correctly in the objects. The results of the simulation of information processing in the system of video monitoring of mobile objects on a real video signal. As a result of the proposed algorithm is achieved by efficient allocation of moving objects in their natural colors without uninformative background. Using this algorithm, for example, in video surveillance systems, will significantly reduce the amount of data that will be stored in the video archive and will simplify the perception of information by the operator and reduce its psychological burden. The results of the algorithm in the future can be used for high-level analysis of mobile objects, in particular, identifying objects and determining their trajectories.

1. N. A. Obukhov, Detection and tracking of moving objects by comparing blocks // Information and Control Systems. 2004. No. 1. P. 30–37. 2. A. Neeti, A survey of Techniques for Human Detection from Video // University of Maryland, Technical report. — 2005. 3. M. Piccardi, Background Subtraction Techniques: A Review IEEE SMC / ICSMC, vol. 4, pp. 3099-3104, 2004. 4. P. Kharebov, Novikov problems selecting objects in a compressed image stream. Proceedings of Graphicon, 2009. 5. 6. R. V. Shestov, A. V. Tamyarov, Brief description of modern methods of morphological processing halftone images among Matlab // Herald VUiT. — 2013. — No. 2. P. 91–98.