Crowd density statistical method and device based on key point detection, equipment and medium

The embodiment of the invention discloses a crowd density statistical method and device based on key point detection, equipment and a medium. The method comprises the following steps: acquiring image data; processing the image data to obtain a sample image; inputting the sample graph into a crowd density statistical model for head probability prediction to obtain a head probability graph; and calculating the crowd number according to the head probability graph. The method not only depends on head key points, but also depends on key points of other parts of the human body for recognition, model supervision information is increased, the representation capability of the human body density feature map is improved, the model training convergence speed is increased, crowd density prediction is facilitated, and the problem that people counting is missing due to partial shielding of the human body parts is solved. In addition, the accuracy of predicting the crowd density map by the model is improved through multi-fea