炸药量推算在爆炸现场分析中具有重要作用。利用炸坑痕迹的3个表观量化特征参量,建立了基于广义回归神经网络的炸药量推算方法,并以野外爆炸实验数据为基础,进行炸药量推算结果与实际值的比较。结果显示,此方法在根据土介质炸坑痕迹推算炸药量方面能够达到比较准确的推算效果,尤其在软粘土及中硬粘土介质中,对炸药量的推算具有较高的准确性和可靠性。 The calculation of explosives plays an important role in the analysis of the explosion scene. Based on the three apparent quantitative parameters of fried pits, a method for estimating the amount of explosives based on generalized regression neural network is established. Based on the field explosion experimental data, the calculation results of explosives are compared with the actual values. The results show that this method can reach more accurate projections based on the explosives trace of soil medium, especially in soft clay and medium-hard clay medium, which has high accuracy and reliability in estimating the amount of explosives.