An application of artificial intelligence for boosting digital transformation in the field of forest management

Abstract

Digital transformation has been strongly affecting many factors of different fields and is a crucial tool to enable the fast and sustainable development of economy and the modern society. This paper focuses on studying and applying artificial intelligence, specifically its sub-domain in deep learning, to create a case study of digital transformation in the area of forest management with a particular emphasis on solving the problem of animal and plant recognition. Specifically, the paper proposes using the MobileNetV3 as the backbone network because of its advantages in efficiency and accuracy. Following that, optimized learning techniques such as soft labeling, data augmentation, and transfer learning were used to improve generality and performance. Experimental results showed that the model performs well in terms of both recognition accuracy and inference time in comparison with other methods. Finally, we have developed an application on mobile platforms (iOS and Android) and the deployment test showed promising performance.