Method for predicting resource abundance of acoustic scattering layer based on NASC data optimization and model optimization, and device

Abstract

Disclosed. are a method for predicting resource abundance of an acoustic scattering layer based on nautical area scattering coefficient (NASC) data optimization and. model optimization, and a device. The method includes: determining an effective sampling space of each sea area on the basis of an effective sampling depth in a vertical direction and an effective sampling interval in a horizontal direction of each sea area, and performing secondary sampling to obtain a sample data set of each sea area; performing cluster division on sample data of all sea areas by an optimized clustering algorithm to obtain a standard acoustic scattering layer NASC data set; training a PSO—XGBoost optimization model, and predicting the resource abundance by using the trained PSO—XGBoost. optimization. model; and. optimizing' predicted NASC values by a Kriging interpolation method to obtain a resource distribution map of the acoustic scattering layers in an entire sea area.