The invention discloses a micro-seismic first arrival intelligent
pickup method,
system and device based on an FFPN model, and a storage medium, and belongs to the field of
exploration geophysics, and the model provided by the invention retains the multi-scale
feature fusion capability of an original FPN
network model, is suitable for a first arrival intelligent
pickup problem, and can achieve a more refined first arrival output result. According to the
network model, output of two scale features is selected as an optimization condition of the model, finer first-arrival
pickup output is calculated through low-level features with rich detail features, dichotomous output is calculated through high-level features with a larger
receptive field, and the output loss of the two parts is calculated. And back propagation is carried out in the training process to optimize the
network model, so that a more accurate dichotomy result (a first arrival picking result) is obtained. Moreover, the point-aware
loss function proposed by the model can calculate the loss of low-level feature output, and the low-level features are optimized to achieve the effect of optimizing high-level feature output.