Automatic pedestrian detection and monitoring system based on Deep Learning


Monitoring plays an important role in security and inspections, but it is also a very tedious task. The emergence of deep learning has liberated humans from this task to some extent. This project builds a simple and effective monitoring system based on the goal detection of deep learning, which can automate the flow statistics and pedestrian detection.

This system is based on the Apache2.0 protocol open source, please strictly abide by the open source agreement.

0x00 Introduction

The system consists of the following three sub-projects:

The overall framework is shown below:

0x01 Server Deployment

1.Server configuration requirements

Configuration Basic requirements
OS Ubuntu 16.04 x64
CPU Main frequency 2.0GHz or more
RAM 8G or more
GPU NVIDIA GTX1080 or more
Network The server IP address needs to be the public IP address.

2.Pedestrian detection system based on TensorFlow platform

The system relies on the following:

Dependency Installation method
Python3.5 Skip
pip Skip
TensorFlow-1.11.0-GPU Skip
Python version - OpenCV Skip
requests pip3 install requests
frozen_inference_graph.pb Download Link
Nginx with RTMP Installation Process

How to run the system:

3.Push flow system based on Android platform

How to run the system:

4.Display system based on SSM (SpringMVC+Spring+Mybatis) Internet lightweight framework

The system relies on the following

Dependency Installation method
JDK-1.8.0 Skip
Apache-Tomcat-9.0.12 Skip
Maven Skip
Mysql Need to configure remote access rights

How to run the system:

0x02 Project Display

0x03 About