Population density estimation method based on convolutional neural network

Abstract

The invention discloses a population density estimation method based on the convolutional neural network. The method comprises the steps that S1) a training sample set is established; S2) a convolutional neural network model based on Mixed-Pooling is constructed; S3) the convolutional neural network model is trained, a random gradient descent method is used to carry out iteration training on the convolutional neural network model, the gradient and the value of a loss function are detected every time after iteration, optimal solutions weight values W and bias values b in the network structure are obtained, and an optimal convolutional neural network model of training is obtained after multiple times of iteration; and S4) the population densities of the convolutional neural network classified models, obtained in the step S3), about far and near subareas are estimated and detected, and the population density of the whole area is estimated according to a new detection and classification strategy. According to the invention, the problems including background interference and pedestrian shielding in a complex scene are overcome, and the population density of the scene is estimated accurately.

Description

technical field [0001] The present invention relates to a crowd density estimation method, in particular to a crowd density estimation method based on a Mixed-Pooling convolutional neural network. Background technique [0002] In recent years, with the rapid development of the economy and the gradual improvement of people's living standards, more and more people will choose to travel during holidays, resulting in a dramatic increase in the number of tourists in various scenic spots. The hidden dangers are becoming more and more obvious, and there are more and more safety accidents. Therefore, how to use technologies such as computer vision to intelligently monitor crowds and make early warnings in a timely manner, and to take effective measures, is of great significance for ensuring social stability and the safety of people's lives and property. At present, there are mainly two methods for crowd density estimation: [0003] 1. Method based on pedestrian detection: At prese...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is to provide a method for estimating crowd density based on convolutional neural network to overcome the background interference of complex scenes and pedestrian occlusion for the deficiencies of the existing technology, and then realize the estimation of crowd density in the scene. an accurate estimate of

Images