❑ The template matching technique could be implemented in IR video sequence to recognize the given target, ❑ In a video sequence with many objects, we can still use a template object in a frame of the video to recognize the object in the other frames, ❑ With selecting the size of the template image is nearly identical to the size of the target and a size of sub-image ƒ more than 2 or 3 times the size of template image g, will give us a result with the high accuracy and processing time is also acceptable, ❑ To ensure more accurate, we should open up a new solution for taking the updated template image g after a period of time to avoid a decrease of correlation coefficient., ❑ Firstly, this algorithm is not capable of identifying targets in case its characteristics are changed, for example, obscured by other objects, masked by terrain conditions, ❑ Secondly, the algorithm cannot detect objects at long distances in comparison with background subtraction algorithm, ❑ In this paper we used maxcc ≥ 0.8. The correlation coefficient value is often chosen on dependence on distance, movement direction and size of template image.
Using template matching for object recognition in infrared video sequences
01.09.2015
10598335 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Using Template Matching Technique for Object Detection in Infrared Images
British Library Conference Proceedings | 2014
|Object recognition by multiresolutional template matching [2606-29]
British Library Conference Proceedings | 1995
|Invariant object recognition using a neural template classifier
British Library Online Contents | 1996
|