env.yml: Conda environment to run scripts
plot_object_detection.py: Run the images through the object detection model and spits out an xml file containing the results
pedestrian_detection.py: Parse xml files and track people in the photo using re-id
plot_lines.py: Using database information to trace trajectories and write them to an image (hourly and daily)
SagePedestrian.sql: Sql script that creates the tables for the pedestrian detection database
requirements.txt: Environment creation for non Conda environments
Useful Sources:
Tutorial to create conda env: https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#activating-an-environment
Open CV Image Writing: https://docs.opencv.org/3.4/d4/da8/group__imgcodecs.html (Python is below c++)
Open CV Line Drawing Documentation: https://docs.opencv.org/4.x/dc/da5/tutorial_py_drawing_functions.html
Object Detection Tutorial: https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/auto_examples/plot_object_detection_saved_model.html
Deep Re-id: https://github.com/KaiyangZhou/deep-person-reid
Cronjob Tutorial: https://www.hostinger.com/tutorials/cron-job
Machine Learning Tutorial: https://www.youtube.com/playlist?list=PLQY2H8rRoyvz_anznBg6y3VhuSMcpN9oe
Node.js and NVM: https://ostechnix.com/install-node-js-linux/
STEPS TO GETTING THE CODE UP AND RUNNING (on hartley):
- Clone this github repo
- Create and activate a new conda environment using the provided env.yml file (yml file may not include all packages, just run code and install packages as necessary)
- Clone the deep-person-reid repo into this cloned repository. You shouldn't need to install it, make sure it's in the directory where you are running the script.
- Make sure that the .xml files for the day you run exist in this directory /raid/AoT/image_label_xmls/
- To run the pedestrian_detection script STANDALONE, follow this format: "python pedestrian_detection.py [hour_min] [hour_max] [date1, date2, ...]"
- Where hours are optional, dateN = yyyy/mm/dd
- Otherwise running plot_object detection will call pedestrian_detection and plot lines.
- If not running pedestrian detection alone, use python plot_object_detection.py, which will use current date.
- The output of the pedestrian_detection.py script will be found in /raid/AoT/image_label_xmls/crosswalk_detections/(ran_date)
TROUBLESHOOTING)
-
If the code refuses to use the gpu, try adding this to your .bashrc script in your home directory export CUDA_VISIBLE_DEVICES=0 export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/[user]/.conda/envs/tf_gpu/lib export PYTHONPATH=$PYTHONPATH:/home/[user]/work/models
-
"ModuleNotFoundError: No module named 'torchreid'" -> Double check the deep-reid path at the top of pedestrian_detection.py if you get this error, make sure it points to the deep-reid directory
- For any "ModuleNotFoundError", "pip install [module_name]" will usually do the trick.
- In the case of cv2, try: pip install opencv-python
-
If you are running a specified day on the plot_object_detection.py script and want it to run in the background, follow the steps commented at the top of that file
-
Plot Lines can be run standalone but you must change the date at the bottom of the script in main
Helpful MySQL commands:
Select all the people in a specific day with their coordinates - select PERMAID,XCOORD,YCOORD from Coordinate where DATE like 'yyyy-mm-dd%';
- to get specific times use YYYY-MM-DD HH:MM:SS
Find ID's of the people who have crossed the road - select PERMAID, USEROAD from Person where USEROAD = 1;
- replace USEROAD with USECROSSWALK for crosswalk
Find out what days are in the database - select DATE from Frame where DATE like 'yyyy-mm-dd%';
Find the total number of people who used the crosswalk - select count(PERMAID) from Person where USECROSSWALK = 1;
Find the number of people who used the crosswalk in a certain day - select count(a.PERMAID) from ( select PERMAID from Person where USECROSSWALK = 1 INTERSECT select PERMAID from Contains where DATE LIKE '2022-05-09%' ) as a;
Table Names : Contains, Coorindate, Frame, Person
Person
+--------------+-------------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+--------------+-------------+------+-----+---------+-------+
| PERMAID | int(11) | NO | PRI | NULL | |
| DAYID | int(11) | NO | | NULL | |
| USECROSSWALK | smallint(6) | YES | | 0 | |
| USEROAD | smallint(6) | YES | | 0 | |
| NS | smallint(6) | YES | | NULL | |
| EW | smallint(6) | YES | | NULL | |
+--------------+-------------+------+-----+---------+-------+
Coordinate
+---------+----------+------+-----+---------+----------------+
| Field | Type | Null | Key | Default | Extra |
+---------+----------+------+-----+---------+----------------+
| TOTAL | int(11) | NO | PRI | NULL | auto_increment |
| PERMAID | int(11) | NO | MUL | NULL | |
| DATE | datetime | NO | MUL | NULL | |
| XCOORD | int(11) | YES | | 0 | |
| YCOORD | int(11) | YES | | 0 | |
+---------+----------+------+-----+---------+----------------+
Frame
+---------+-----------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+---------+-----------+------+-----+---------+-------+
| DATE | datetime | NO | PRI | NULL | |
| PATH | char(128) | NO | | NULL | |
| FRAMEID | int(11) | NO | | NULL | |
+---------+-----------+------+-----+---------+-------+
Contains
+---------+----------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+---------+----------+------+-----+---------+-------+
| PERMAID | int(11) | NO | PRI | NULL | |
| DATE | datetime | NO | PRI | NULL | |
+---------+----------+------+-----+---------+-------+
This material is based upon work supported by the National Science Foundation under Grant No. OAC 1935984.