Real-Time Monitoring of Illegal Logging Events Using Intelligent Acoustic Sensors Nodes

The usage of sensor networks for monitoring and assessing various facets of our everyday life has grown significantly. Illegal logging activity identification is critical in today's forest environment to conserve habitat. Acoustic signals are a significant tool in monitoring illegal logging activities over wireless acoustic sensor networks (WASNs). However, the continuous monitoring of acoustic signals requires a high computational demand that prevents real-time computations within the nodes. This article presents an efficient methodology based on acoustic signal processing and its classification for real-time remote monitoring and detection of illegal logging activities. The proposed method consists of a short-time Hartley transform (STHT)-based spectrogram for extracting the low-level audio features (LLFs) from the acoustic signals. Subsequently, these extracted features are classified into five acoustic classes using k-nearest neighbor (kNN), decision tree (DT), random forest (RF), adaptive boosting (Ada-Boost), and support vector machines (SVMs) classifiers. The efficient combination of feature representation and classification method is implemented on a 32-bit microprocessor platform, interfaced with a LoRa wireless module to develop an intelligent acoustic sensor nodes (IASN) system, which is validated by regenerating various acoustic signals along with forest ambience using loudspeakers in an experimental laboratory setup. Once an illegal logging activity is detected, the edge node generates an alert message and sends it to the mobile phone through the gateway. The accuracy of the proposed system is 96.61%, along with sensitivity, specificity, and F-score of 96.20%, 99.14%, and 96.07%, respectively.