Publication: Development of algorithm to detect presence of animal in digital image
Loading...
Date
2012-06-01
Authors
Koik, Boon Tatt
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Animal detection can be used in wide range of applications especially intrusion detection of animal for security reason to ensure the safety of human life and assets. Therefore, developments of an algorithm to detect the presence of animal become a topic in various researches. The main purpose of this research work is to process multiple static image framesm which act as a stream of video that been chopped into multiple images. In this research, an image processing method is proposed using multiple image frames comparison approach based on background subtractions. Every pixel in the multiple images subtract with other images, followed by predefined threshold filtration. The purpose of the threshold filtration is to filter the noises of natural image is performed to obtain an initial, rough output to be further processed. Multiple images pixels’ data in Boolean expression are compared to reduce false positive detection of animal is performed. To reduce the noise further and for a more complete detection of animal, connected component and region labelling operation being implemented to group and classify detected area region, followed by area threshold’s operation to further eliminate unwanted anomalies in image frames. In this way, the detected animal in images will be more obvious in the result’s images. At last, the detected animal in images will be marked with rectangle frame of red colour to be further processed. Since the proposed method can work various types of input data, its potential to be applied in real life animal intrusion detection scenario is very high.