Abstract
Surveillance video is used for security purpose in our daily life in various places. It is used to observe the unusual activity that is taking place around us. Today in most of the shop owners have CCTV cameras to record, the uncertain activities and even it is used in houses in remote places. A system must be smart enough to detect. This paper uses SIFT and SURF algorithm for detection. Image registration is a development in which more than two images from various imaging equipment are reserved at various angles and at various times from the identical prospect and geometrically aligned for further exploration. Data may be from different sensors, CCTV taken at different times, depths, or perspective. Feature-DetectorDescriptor plays a vital role in feature matching application for selection of feature; this paper presents a comparative analysis of SIFT, SURF, algorithms. Experiments have been conducted on a wide range of images taken from datasets. A quantitative comparison is presented. This paper gives an useful ideas for making important decisions and it also helps in providing a smart security system.
Highlights
Image registration applications are Real Time Alerts which includes user defined alerts, Generic Alerts, Classic Specific Alerts and Behavioral Alerts for Video Processing
Features are detected for images using Scale Invariant Feature Transform (SIFT) and Speeded Up Robust Features (SURF) algorithm
This paper provides a complete assessment of SIFT and SURF based on feature-detector-descriptors
Summary
Image registration applications are Real Time Alerts which includes user defined alerts, Generic Alerts, Classic Specific Alerts and Behavioral Alerts for Video Processing. Developing an optimized method that suits for all users as there are wide ranges of application due to images geometric radiometric distortions and noise disturbance, data characteristic and accuracy threshold level are still in process. The significance of image registration techniques develops with developing number of different types of data acquisition devices and their increasing availability. Objects needed for image registration can be aligned that can be translated and rotated to one another
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
