A Video Image-Based Approach for Crowd Counting and Future Traffic Prediction

Abstract

The invention discloses a method for crowd counting and future human flow prediction based on video images. The present invention: 1. Select a video image data set with annotation information, perform Gaussian function processing according to the annotation of the head position, and generate a true value density map; 2. Input the video frame into the built MPDC model to extract the feature map, and map it as Crowd estimated density map (DE). 3. Input the obtained DE stacking frames into the constructed Bi‑ConvLSTM network, predict the crowd prediction density map at time T+1, and estimate the number of pedestrians at time T+1. The present invention adopts the convolution network based on multi-scale pyramid holes and the Bi-ConvLSTM network based on residual connection, uses continuous video frames to generate crowd estimation density map, and further predicts crowd prediction density map of future frames, and counts the crowd. quantity. The present invention is aimed at the prediction of continuous video images, and is a brand-new method, which can not only obtain the real-time crowd density map and the number of people, but also predict the crowd density map and the flow of people in future frames.

Description

technical field [0001] The invention belongs to the field of crowd image processing in computer vision, and more particularly relates to a method for crowd counting and future crowd flow prediction based on video images. Background technique [0002] Crowd counting is to count the number of people in a picture or video sequence. Crowd counting and prediction are of great significance to public safety management, regional spatial planning, and information resource acquisition. It can better monitor and divert crowds in public places, and provide a basis for rational scheduling of personnel, rational planning of routes, rational dredging of people flow, and site selection of buildings. [0003] Existing crowd counting methods can be divided into three categories: crowd counting methods based on detection, regression, and density map estimation. Detection-based methods are suitable for larger and more sparse target scenes. However, the method of regressing the number of crow...

AI Technical Summary

Problems solved by technology

[0005] However, in complex backgrounds and scenes where high-density crowds gather, due to the influence of interference factors such as target occlusion caused by high overlap between people, perspective perspective, scale change, and uneven density distribution, the counting of crowds and the impact of density information There is great resistance to obtaining

However, the existing methods for generating density maps based on deep learning usually use multi-column convolution layers and large convolution kernels for feature extraction of multi-scale images, thus generating a large number of parameters and increasing the difficulty of network training.

In addition, there is no research on crowd flow prediction algorithm based on crowd video images.

Images