Advanced Journal of Environmental Science and Technology

ISSN 2756-3251

Advanced in Environmental Science and Technology | Vol. 17 (5), pp. 281-293, May, 2026

Research Article

Acoustic Tomography and Deep Learning Acoustic Event Classification for Real-Time Monitoring of Illegal Logging and Forest Degradation in Tropical Rainforests

Gabriel Santos¹*, Jean-Baptiste Nkosi², Mei-Ling Tan³

¹Department of Forest Sciences, Federal University of Amazonas, Manaus, Brazil
²Department of Environmental Science, University of Kinshasa, Kinshasa, Democratic Republic of the Congo
³Asian School of the Environment, Nanyang Technological University, Singapore, Singapore

Corresponding Author: Gabriel Santos — [email protected]

Abstract

Optical satellite monitoring of tropical rainforests frequently suffers from persistent cloud cover and delayed detection of selective logging beneath intact upper canopies. This research presents a solar-powered wireless acoustic sensor network utilizing multi-node time-difference-of-arrival (TDOA) acoustic tomography paired with a lightweight convolutional recurrent neural network (CRNN) for real-time edge classification of anthropogenic forest disturbances. Deployed across 15,000 hectares of lowland rainforest, the edge-AI nodes classified chainsaw operation, heavy machinery transit, and tree-felling impacts with 96.8% precision while filtering out bioacoustic noise from insects, birds, and tropical rainstorms. Acoustic triangulation localized illegal logging events within a 35-meter radius and transmitted low-bandwidth LoRaWAN alerts to forest rangers within 90 seconds of onset, enabling rapid ground intervention prior to timber extraction.

Keywords: ecoacoustics; deep learning; illegal logging detection; tropical forest conservation; wireless sensor networks; acoustic tomography

Manuscript Timeline: Received 07 April, 2026; Revised 06 May, 2026; Accepted 20 May, 2026; Published 28 May, 2026.

Citation: Santos, G., Nkosi, J.-B., and Tan, M.-L. (2026). Acoustic Tomography and Deep Learning Acoustic Event Classification for Real-Time Monitoring of Illegal Logging and Forest Degradation in Tropical Rainforests. Advanced in Environmental Science and Technology, 17(5), pp. 281-293.