Detection and Counting of Animals in CameraTrap Images using Faster R-Cnn and Data Augmentation Techniques

Panawè Touh,Mukesh Sharma

doi:10.35940/ijeat.e9925.069520

Abstract

Camera traps are used to recover images of animals in their habitats to help in the conservation of fauna. Millions of images are captured by camera traps and extracting information from these data delays and consumes enough resources so sometimes millions of images cannot be used due to lack of resources. That is why researchers have proposed solution approaches using Convolutional Neural Networks (CNNs) and object detection models to be able to automate the retrieval of information from these images. We used Faster R-CNN and data augmentation techniques on Gold Standard Snapshot Serengeti Dataset to detect animals in images and count them. The performances of the two models (the one trained on the original dataset and the one trained on the augmented dataset) were compared to show the importance of having more data for this task. Using the augmented dataset, we trained our model which reached an accuracy of 98.26% for classification of the proposed regions, an accuracy of 79.55% for counting the species present on the images and a mAP of 95.3%. For future work, the model can be trained to recognize the actions and characteristics of animals and tuned to be more efficient for counting task.

Highlights

It is widely assumed that with the use of camera traps and especially with their proliferation in many projects which can be counted per hundred, we have better understood the natural environment and species with the data collected [1]
Gold Standard Snapshot Serengeti Bounding Box Coordinates [7] is a dataset which contains 4010 images labelled by experts in the field and which have been annotated by Stefan Schneider et al
We investigated whether using data augmentation techniques on the original dataset containing images of animals captured by camera traps improves the detection performance of animal species in these images using an object detection model (Faster R-Convolutional Neural Networks (CNNs) [4])

Summary

Introduction

It is widely assumed that with the use of camera traps and especially with their proliferation in many projects which can be counted per hundred, we have better understood the natural environment and species with the data collected [1]. Even with almost perfect efficiency, camera traps pose a major problem for researchers because they generate a lot of images so significant resources such as time and manpower are used to analyse all of these images in order to extract information necessary for studies. With the development of very efficient techniques and models in the field of deep learning, automatic information extraction solutions have been proposed to overcome this problem. Deep learning models gave the ability to computers to analyse and understand digital images [2].Identifying and collecting information for each object in images is more beneficial and provides more information than classifying images into a class [3].

Objectives

Results

Conclusion