A Real-Time Through-Wall Detection Based on Support Vector Machine

Abstract

In this paper, a real-time approach for solving the throughwall imaging is proposed based on support vector machine. The proposed technique avoids the consideration and computation of the effect of the walls, such as refraction and speed change, which is a primary difficulty and must be computed using the conventional methods. Numerical results show that once the training phase is completed, this technique only needs computational time in an order of seconds to locate the target with small errors. Noiseless and noisy measurements are discussed, making the proposed technique more realistic. Numerical experiments also show how the method is capable, when noise contamination is considered, to locate targets in throughwall scenes. This validates the effectiveness of the proposed approach.