ISSN 2348-1196 (print) International Journal of Computer Science and Information Technology Research ISSN 2348-120X (online) Vol. 10, Issue 4, pp: (10-16), Month: October - December 2022, Available at: www.researchpublish.com
Detection of Endemic Sri Lankan Birds Using AI-Based Software 1
K.A.S.H Kulathunga, 2S.A.M Piyabhashitha, 3S.A.D.H.A Jayatilake, 4Jeewaka Perera, 5 Dhammika Silva 1, 2,3,4,5
Faculty of Computing, Sri Lanka Institute of Information Technology Malabe, Sri Lanka DOI: https://doi.org/10.5281/zenodo.7234060
Published Date: 21-October-2022
Abstract: Photographers frequently encounter a variety of issues when taking wildlife photographs. Many local and foreign wildlife photographers struggle with the lack of an efficient tool for detecting and discovering endemic Sri Lankan birds. A mobile application called "Ceylon Birds" was developed as a solution for these issues. It uses birds’ images, voices, and habitats to identify the birds. This mobile application will access the device’s camera, recorder, and Global Positioning System to accurately identify the bird, habitats and provide the bird’s details to the photographer. The concepts of Machine Learning, Natural Language Processing, and Neural Networks are used for this application. The information supplied by the Wildlife Officers, experts in this sector, was used to develop this application. The main goals of the suggested solution are to locate the regions where the majority of birds are present during a relevant time period and to clearly identify endemic birds by their physical characteristics and tones of voice. The trained Machine Learning models have achieved the accuracy of 92%, 90%, and 88% for the voice detection model, image identification model, and location clustering model respectively. After testing this “Ceylon Birds” mobile application among wildlife photographers, we have received positive feedback from them. Keywords: image identification, voice detection, location clustering, machine learning.
I. INTRODUCTION One of the most standard photographic criteria among many photographers is wild photography. According to data gathered by a survey, wild photography criteria have ranked high among many photographers whereas indoor function photography, outdoor function photography, and object photography ranked as more typical photography criteria. Participants in this survey stated that elegance blended with the diversity of nature that is beautifully caught by the lenses always tends to attract their attention. In Sri Lanka, there are so many bird photographers, but most of them did not identify the birds by their appearance or vocals. Some of the villagers are known about some of those birds. But the problem is many villagers also cannot find the endemic Sri Lankan birds. There is no exact knowledge about endemic birds. Therefore not only the local photographers, foreign photographers also face major problems in the endemic bird identification process. There are many applications available for bird identification. But the issue is those applications did not support for identify endemic Sri Lankan birds. And also those applications did not develop as online image and vocals upload format. Hence to find the bird’s details, photographers need to scroll and find out. It is much harder to do those things at that stage. Therefore it is really important to implement an AI (Artificial Intelligence) based application for wild photographers to detect endemic Sri Lankan birds. The main objectives of the ‘Ceylon Birds’ application are detecting whether the bird is endemic or not and displaying the birds’ details. If the bird is covered by any object system it will preprocess the captured image and display the relevant bird’s name. In case there is too much noise in the environment system will filter out those background noises and display the relevant birds’ names. In both cases, if the endemic birds cannot be identified, the application will provide the feature to identify the birds by asking questions by using their habitats and appearances. And also habitats of the
Page | 10 Research Publish Journals