The invention relates to the technical field of concrete production, in particular to a fine aggregate
moisture content detection method and device based on a depth image and a
microwave frequency spectrum, and the method comprises the following steps: acquiring material thickness curve data and a
microwave frequency spectrum curve of fine aggregate by combining an RGB-D
imaging technology and a
microwave sensing technology; the material thickness curve data and the
microwave frequency spectrum curve are input into a trained multi-mode
moisture content prediction model, the trained multi-mode
moisture content prediction model comprises a material thickness
feature extraction network, a
microwave frequency spectrum
feature extraction network and a
feature fusion network, the material thickness curve data passes through the material thickness
feature extraction network, material thickness features are extracted, and a
microwave frequency spectrum feature extraction network is obtained; the microwave spectrum curve passes through a microwave spectrum feature extraction network, microwave spectrum features are extracted, material thickness features and the microwave spectrum features are input into a
feature fusion network, and a predicted value of the moisture content is obtained through prediction. The problems that in the prior art, fine aggregate moisture content detection is not accurate, and the
delay performance is high are solved, material thickness fluctuation is effectively filtered out, and the moisture content
measurement precision and stability are improved.