Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Crowd counting method for subway carriage scene

A crowd counting and compartment technology, applied in the computer field, can solve problems such as large changes in crowd scale, inability of neural networks to adapt and identify crowd characteristics, and affect the accuracy of counting

Pending Publication Date: 2021-04-09
苏州玖合智能科技有限公司
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] To achieve accurate crowd counting in subway cars, the main difficulty currently faced is that the crowd scale changes too much due to the large depth of field, and the conventional neural network cannot adapt to and recognize the characteristics of crowds of various scales, thus affecting the accuracy of counting

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Crowd counting method for subway carriage scene
  • Crowd counting method for subway carriage scene
  • Crowd counting method for subway carriage scene

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] Below in conjunction with accompanying drawing and specific implementation mode this scheme is further described:

[0036] 1. Program overview:

[0037] 1. Pre-training part

[0038] 1.1. Data preprocessing:

[0039] refer to figure 1 , because the data captured by the camera is a video file with severe distortion, the different degrees of distortion of the crowd seriously affect the recognition of the head features by the network, therefore, the data preprocessing part is as follows Figure 1 As shown in the figure, it includes three main steps of intercepting a single frame picture, distortion processing, and image enhancement.

[0040] 1.2. Network pre-training

[0041] refer to figure 2 with 4 For the situation where the scale changes are too large, first cut the picture into upper and lower parts, and reduce the scale change of each picture to a controllable range, so as to improve the accuracy of crowd counting. Send large-scale head-scale pictures to the ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a crowd counting method for a subway carriage scene, and the method comprises the steps: 1, employing a wide-angle camera to obtain image data in a carriage, and processing the data; and 2, sending the data to a crowd feature recognition network for processing, and performing crowd counting. In the step 1, pictures of a real-time video are captured according to time, correction and enhancement are carried out, and then real-time data are sent to the step 2 to be processed; in the step 2, for real-time data, whether the real-time data is sparse crowd data or dense crowd data is judged; and the crowd feature recognition network uses the weight A and the weight B for sparse and dense crowd data counting respectively.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to an application of computer vision technology in crowd counting. Background technique [0002] The main task of crowd counting is to identify crowd features from images and accurately calculate the number of crowds in the image. Early crowd counting was divided into detection-based and regression-based methods. In detection-based methods, a sliding window detector is used to detect crowds in the scene and count the corresponding number of people. Detection-based methods are mainly divided into two categories, one is whole-body-based detection, and the other is part-body-based detection. The detection method based on the whole, for example, the typical traditional method, mainly trains a classifier, and uses the wavelet, HOG, edge and other features extracted from the whole body of the pedestrian to detect pedestrians. Learning algorithms mainly include methods such as SVM, bo...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F21/62G06K9/62G06N3/04G06N3/08
CPCG06F21/6227G06N3/08G06N3/045G06F18/214Y02T10/40
Inventor 田青唐绍鹏
Owner 苏州玖合智能科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products