Real-time object detection using dynamic principal component analysis

In this work, we contribute to the real-time detection of buried objects, with special emphasis on explosive ones, using the ground penetrating radar (GPR). When the buried objects have explosive substance, the moment of detection becomes vital. Therefore, we start the detection process right after the very first GPR signals begin to return from the buried objects. For this purpose, we adopted the studies focusing on the online process monitoring methods using principal component analysis (PCA), and adapted them to the dynamic conditions of the ground. Different objects with varying dielectric properties are buried in the test environment and used for the evaluation of the proposed method. With the observed results, it is validated that, the proposed method is employable towards the real-time object detection.