elevator-214 / elevator-monitor

Intelligent monitoring of escalator.Function including traffic statistics,passenger retention detection and large object retention detection in escalator floor board. As well as human keypoints extraction and tracking in elevator.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Introduction

Use three cameras to monitor some behaviors on the walking elevator.

Including:

  • Traffic statistics
  • Passenger&Large Object Retention Detection on escalator floor board
  • Human Keypoints Extraction and Tracking in elevator
  • Detection of object sticking out of handrail

This project is implemented in C++ and uses QT for interface interaction. Just a student project made by a total newb ,so feel free to ask and don’t expect too much. Test on Jetson Xavier. It should also work on ubuntu x86/64.

TO DO LIST

  • [✔] Detect the heads by YOLOv3-tiny
  • [✔] Tracking the human heads by Kalman Filter and Hungarian match
  • [✔] Human retention detection on escalator floor board
  • [✔] Passenger flow statistics
  • [✔] Large items retention detection by KNN+kd-tree and Hungarian match
  • [✔] Detecting object sticking out of handrail by GMM
  • [✔] Human keypoints extraction by Openpose+TensorRT
  • [✔] Human keypoints tracking

Pre-installation required

CUDA、cudnn、TensorRT6、Boost、QT5、OpenCV4.2

Install the Project

cd elevator-monitor
mkdir build
cd build
cmake .. && make
cd .. && bash ./run.sh

Some models & files required

BaiduYunPan url: https://pan.baidu.com/s/1HLDYnXif759s__yVIvGd_Q

Extraction code: 1pxh

usage:Download and move models & files to build/

Reference

https://github.com/gishi523/kd-tree

https://github.com/saebyn/munkres-cpp

https://github.com/zerollzeng/tensorrt-zoo

About

Intelligent monitoring of escalator.Function including traffic statistics,passenger retention detection and large object retention detection in escalator floor board. As well as human keypoints extraction and tracking in elevator.


Languages

Language:C++ 76.9%Language:Python 12.1%Language:Cuda 7.8%Language:CMake 2.8%Language:C 0.3%Language:Shell 0.1%Language:Objective-C 0.0%