The purpose of this project is to counting the people passing through the bus by RGB-D camera accurately.

This project starts from 2015.01 to 2017.05.

designed a software and hardware system

We have designed a software and hardware system.

The software system includes: calibration, one-camera counting, double camera counting, recording, auto testing. The hardware system.

The hardware system includes: designed RGB-D camera, redesigned mini-computer. Besides, we also integrated our algorithm into the camera.

published a dataset

[github]

We have published a dataset, called People counting dataset (PCDS). It is divided into 4 sub-categories: N+C+, N+C-, N-C+, N-C- according to the sunlight and crowed, as shown in the following table.

	N+C+	N+C-	N-C+	N-C-
sunlight	strong	strong	weak	weak
crowed	yes	no	yes	no
entering	937	616	5427	2704
exiting	1149	668	6647	2760
total	2086	1284	12074	5464

There are total up to 5,464 videos in this dataset. The detail number of people entering and exiting the bus in each category can be referred from the following table.

published a paper

[pdf ]

We have written a paper to clarify our algorithm. This paper’s name is “Benchmark data and method for real-time people counting in cluttered scenes using depth sensors”.

Besides, two patent has been published, as follows:

• 宋焕生,孙士杰,张朝阳,刘瑞芝,张文涛,崔华,李钢,李怀宇,张向清,李莹,潘强,王璇,杨燕妮,孟乔,孙亚,李婵. 一种基于轨迹分析的人数统计方法[Z]. CN106600643A: 长安大学,2017.
• 宋焕生,孙士杰,刘瑞芝,张文涛,张朝阳,崔华,李钢,李怀宇,张向清,李莹,陈艳,王璇,杨燕妮,孟乔,潘强,李婵,孙亚. 一种基于RGB#D相机的人数统计方法[Z]. CN106570883A: 长安大学,2017.
• 宋焕生,孙士杰,庞凤兰,刘瑞芝,李倩丽,王璇,张文涛. 一种基于3D轨迹分析的目标检测识别算法[Z]. CN106127137A: 长安大学,2016.
• 宋焕生,孙士杰,刘瑞芝,张朝阳,张文涛,李莹,张向清,李怀宇,崔华,李钢,陈艳,王璇,李婵,孙亚. 一种复杂场景下的人数统计方法[Z]. CN106228560A: 长安大学,2016.
• 宋焕生,孙士杰,张朝阳,刘瑞芝,王璇,陈艳,李怀宇,崔华,张文涛,张向清,李莹,严腾,郑宝峰,张斌. 一种基于三维映射表的三维点云快速重建方法[Z]. CN105913488A: 长安大学,2016.