Pig detection method based on transfer learning and multi-source domain fusion

The invention provides a pig detection method based on transfer learning and multi-source domain fusion, relates to the field of animal husbandry artificial intelligence, constructs an intermediate domain through a high-definition adaptive improved generative adversarial network model, migrates the intermediate domain to a generative domain, and provides a transfer learning evaluation algorithm. Data pictures and the like of other breeding factories are migrated to a target pig farm to strengthen a training data set, a large number of data sets can be obtained only through a small amount of photo data, the image acquisition cost is greatly saved, convenience is provided for large-scale application of artificial intelligence to small and medium-sized farms, and target detection is carried out through a frame-free model to improve the detection precision. The method can be applied to large-scale farm target identification, target tracking can serve as the basis of behavior identification and image segmentation,