Pedestrian volume detection method based on deep learning

The invention discloses a pedestrian volume detection method based on deep learning, and relates to the technical field of artificial intelligence. Compared with a mainstream convolutional neural network, the lightweight neural network model used in the invention decomposes the standard convolution kernel through deep separable convolution, reduces the calculation amount, accelerates the calculation, has excellent performance, and can reduce the size of the model and increase the speed on the premise of maintaining the performance of a traditional model; the defect of insufficient memory caused by an overlarge model is overcome, and the method is suitable for deployment of a mobile terminal or an embedded chip. According to the pedestrian volume detection method under video monitoring, a practical data set is made on the basis of the lightweight network model framework, and the method is suitable for special scenes such as parks and cultural squares with a large number of baby carriages and low pedestrian mov