| Summary: | Illegal logging is a continuous environment issue not just in Malaysia but all over the world. It threatens the environmental sustainability, shrinks state government's income and undermines the right of local community. Furthermore, it can cause deforestation, greenhouse effect and imbalance of the eco-system. Various efforts have been placed to reduce this activity such as drones, satellite and remotes sensing, however, it can be only detected if there are changes in the image received from those technologies which mean deforestation is already happened. Hence, forest surveillance and monitoring, as well as automatic early detection of the intruders, plays important roles in preventing and ending illegal logger activities. Therefore, this paper presents vehicle detection and classification system based on Support Vector Machine (SVM) in Forward Scatter Radar to detect illegal vehicle entering the forest. Measured signals from four different vehicles are used as the input of the system. To achieve high accuracy in classifying the vehicle, certain features are extracted from the spectra of the vehicles and fed as the input of SVM. The result shows that different classes of vehicles can be separated and classified using the proposed classification algorithm. Although, the are some overlapping of data for similar size of vehicles, the separation of data can be further improved by applying hyperplanes. The proposed FSR system has a potential to be used in detecting and identifying vehicle entering the restricted area such as reserved forest henceforth illegal logging activities can be curbed. © 2021 College of Engineering, Universiti Teknologi MARA (UiTM), Malaysia.
|
|---|