Automatic fish feed feeding system based on deep learning

The invention relates to the technical field of fishery breeding equipment, and discloses an automatic fish feed feeding system based on deep learning, which comprises an image data acquisition module, hardware equipment, a wind power feeding machine, a feeding algorithm control module, a front-back end module and a cloud service module, according to the system, automatic feeding is achieved in the mode that timed feeding and algorithm detection are combined, timed feeding is integrated to an algorithm according to the daily manual feeding time of a fishing ground, and then video images shot by a camera are detected in real time through a deep learning algorithm model; classifying according to the intensity of ripples on the water surface in the fishpond, the amount of residual bait and the number of fishes on the water surface, and controlling feeding in three states of strong ingestion, weak ingestion and non-ingestion; the method is high in efficiency, low in cost and more stable, a large amount of manpowe