Open AccessTracking Objects on Detector Response using Extended Kalman Filter
Author(s) -
R Hemavathy,
G Shobha
Publication year - 2020
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.d8713.049420
Subject(s) - computer science , extended kalman filter , artificial intelligence , segmentation , kalman filter , computer vision , frame (networking) , detector , identification (biology) , object (grammar) , tracking (education) , set (abstract data type) , video tracking , filter (signal processing) , object detection , track (disk drive) , pattern recognition (psychology) , psychology , telecommunications , pedagogy , botany , operating system , biology , programming language
Video analytics plays a very important role in identification or detection and tracking of objects, this intern find application in many fields and domains. Novel learning methods or techniques built on Neural Networks requires larger dataset for training the results, the output obtained depends on how well the training is done. The proposed method of Weighted Cumulative Summation (WCS) is an approach based on background modelling to segment the moving objects. This method adapts and tunes the background variations instantaneously as the video frame arrives. The segmentation obtained is compared with other basic methods. The result obtained infers improvements in segmentation and in removal of ghost effect in the video. Extended Kalman Filter (EKF) is used to track the detector response. The responses of the detection from WCS are provided as input to EKF to track the moving object. The results are tabulated and represented in the form of graphs for analysis. The results are compared with three different video datasets and the results are noticeably good. The methods WCS can be used in the applications were data set is not available.