DocumentCode :
3580718
Title :
Visual object tracking using particle clustering
Author :
Pradhana, Harindra Wisnu
Author_Institution :
Dept. of R&D, BIT Inti Teknol., Semarang, Indonesia
fYear :
2014
Firstpage :
119
Lastpage :
123
Abstract :
Computer vision been used to estimate object location relatively from observer on many applications. High definition sensor often used to gain accuracy of the object tracking which resulting high processing complexity. Lower resolution sensor simplifies the process with significant accuracy lost. Particle clustering method estimates the object location by grouping several detection data with certain similarity. Instead of detecting edges and corner on the visual data, this paper uses clustering method to group pixels with certain similarity and measure its element. The cluster measured both height and width to estimate the distance of the object from the observer. New color features introduced in this research promising a better detection approach even with low resolution sensor. The proposed approach successfully provides 30fps image analysis with significant color extraction improvement.
Keywords :
computer vision; feature extraction; image colour analysis; object detection; object tracking; particle filtering (numerical methods); pattern clustering; color extraction improvement; color features; computer vision; detection data; high definition sensor; image analysis; object location; object tracking; particle clustering method; Atmospheric measurements; Cameras; Euclidean distance; Image resolution; Particle measurements; Tracking; Visualization; Particle clustering; bearing-only; clustering; computer vision;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Technology, Computer and Electrical Engineering (ICITACEE), 2014 1st International Conference on
Print_ISBN :
978-1-4799-6431-4
Type :
conf
DOI :
10.1109/ICITACEE.2014.7065726
Filename :
7065726
Link To Document :
بازگشت