Abstract: The blast point detection is a foundation of fire emendation, damage evaluation and opposite firepower point estimation. According to the characteristics of artillery blast point, this paper proposes a new center-surround Haar-like feature which can be calculated quickly. An algorithm for the blast point detection is presented based on Adaboost with the new Haar-like feature. The experiment results show that the proposed algorithm has better performance than the Adaboost algorithm without the new Haar-like feature.