An Approach for the Development of Animal Tracking System

An Approach for the Development of Animal Tracking System

N. Manohar, Y. H. Sharath Kumar, G. Hemantha Kumar
Copyright: © 2018 |Pages: 17
DOI: 10.4018/IJCVIP.2018010102
(Individual Articles)
No Current Special Offers


In this article, the authors propose a system which can identify and track animals. Identification and tracking of animals has got plenty of applications like, avoiding dangerous animal intrusion into residential areas, avoiding animal-vehicle collisions, and behavioral study of animals and so on. Previously, biologists studied videos to detect and identify animals, a time consuming and difficult task. This requires a fully automatic or computer-assisted system to identify and track animals by video. Initially, frames are extracted from the given video. Segmentation is done to the extracted frames using a maximum similarity-based region merging algorithm. Then, the mean shift-based algorithm is used to track the animals. Finally, the animals are classified using Gabor features and a KNN classifier. Experimentation has been conducted on a data set containing more than 150 videos with 15 different classes.
Article Preview

1. Introduction

Nowadays, with the increase in the availability of high quality cameras at a very low price the demand for automated analysis of videos for various applications has also increased. Object tracking is one such application of automated video analysis which is applicable in many tasks like detecting unusual events in automated surveillance, traffic monitoring to control the traffic flow, eye gaze tracking for data input to computers etc. Thus, object tracking has become an important task in the field of computer vision. An interesting application of object tracking is animals tracking system. Tracking of animals places an important role in monitoring animal behavior with the environment and also to study the locomotive behavior of the same. Issues such as low-quality cameras, multiple animals, animals with different pose, different illumination, occlusion, complex background, fast moving nature of animals etc., have made automated animal tracking system a complex job. Also, the usage of high quality cameras increases the size of the data base which makes the process of tracking and detection a time consuming task. All the above factors influence to develop an automatic tracking system for animals.

The process of animal tracking is summarized in the Figure 1. The procedure first executes a file to convert the selected video file into frames. Then segmentation is carried out to obtain the object of our interest by removing the background regions. The segmented frames will be the input for tracking algorithm which tracks the animals in the consecutive frames. After tracking, object detection module carries out matching between the training images and testing image obtained after feature extraction and segmentation in order to recognize object. The output of the module labels the animal detected in the successive frames.

Figure 1.

Flow diagram of animal tracking and detection system


2. Literature Survey

A variety of different approaches have been proposed in the past to deal with visual recognition of animals. We have reviewed here briefly some of the main approaches, focusing in particular on methods that are applicable to segmentation, tracking, identification and classification.

Complete Article List

Search this Journal:
Volume 13: 1 Issue (2023): Forthcoming, Available for Pre-Order
Volume 12: 4 Issues (2022): 1 Released, 3 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 2 Issues (2016)
Volume 5: 2 Issues (2015)
Volume 4: 2 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing