MouseVUER: Video based open-source system for laboratory mouse home-cage monitoring

About

An opne source home-cage design based on Intel® RealSense™ depth cameras.

CAD Files and Assembly

The CAD files and assembly instructions can be found in the CAD_Models folder.

Manuscript data

The images and annotations used in the manuscript can be found here Manuscript_Data folder

Video Acquisition Software

Depth Video Encoding Pipeline

Directions for Installation

Install librealsense with Ubuntu 20.04 (as of 3/28/2023)

This is a guide on how to run the program To install with linux please follow directions here https://github.com/IntelRealSense/librealsense/blob/master/doc/distribution_linux.md

Install python3 modules

sudo apt update && \
sudo apt install python3-pip && \
python3 -m pip install scapy && \
python3 -m pip install pycrypto && \
python3 -m pip install ffmpeg-python && \
python3 -m pip install numpy && \
sudo apt-get install python3-tk

Install Boost, FFmpeg, GLFW3 and others using the following commands:

Type the following into the command line:

sudo apt-get -y install libboost-all-dev && \
sudo apt-get -y install libglfw3 && \
sudo apt-get -y install libglfw3-dev && \
sudo apt-get -y install freeglut3-dev libgl1-mesa-dev libglu1-mesa-dev  && \
sudo apt-get -y install libjsoncpp-dev && \
sudo apt-get -y install ffmpeg && \
sudo apt-get -y install libavcodec-dev libavutil-dev libavformat-dev libswscale-dev

OPTIONAL Install Intel QuickSync

sudo apt install intel-media-va-driver-non-free x264 libva-dev libmfx-dev intel-media-va-driver-non-free x264 intel-media-va-driver-non-free i965-va-driver-shaders && \
export LIBVA_DRIVER_NAME=iHD

More info https://trac.ffmpeg.org/wiki/Hardware/VAAPI

Install OpenCV with the following directions

https://linuxize.com/post/how-to-install-opencv-on-ubuntu-20-04/

If the above doesn't work

from https://docs.opencv.org/4.6.0/d7/d9f/tutorial_linux_install.html

1. Install manually using guide below.

python3 -m pip install pyrealsense2 && \
sudo apt install -y g++ && \
sudo apt install -y cmake && \
sudo apt install -y make && \
sudo apt install -y wget unzip && \
sudo apt install -y git && \
cd ~/ && \
wget -O opencv.zip https://github.com/opencv/opencv/archive/4.x.zip && \
unzip opencv.zip && \
mv opencv-4.x opencv && \
mkdir -p build && cd build && \
cmake -DOPENCV_GENERATE_PKGCONFIG=ON ../opencv && \
make -j5 && \
sudo make install && \
sudo ldconfig -v

2. Then

sudo nano /usr/local/lib/pkgconfig/opencv4.pc

3. And if nothing is there then copy the following to the opencv4.pc file please replace Version field with your specified version number :

# Package Information for pkg-config

prefix=/usr/local
exec_prefix=${prefix}
libdir=${exec_prefix}/lib
includedir=${prefix}/include/opencv4

Name: OpenCV
Description: Open Source Computer Vision Library
Version: 4.7.0 
Libs: -L${exec_prefix}/lib -lopencv_gapi -lopencv_highgui -lopencv_ml -lopencv_objdetect -lopencv_photo -lopencv_stitching -lopencv_video -lopencv_calib3d -lopencv_features2d -lopencv_dnn -lopencv_flann -lopencv_videoio -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
Libs.private: -ldl -lm -lpthread -lrt
Cflags: -I${includedir}

4. Then do

sudo nano ~/.bashrc

5. Now copy the following at the bottom of the .bashrc file

PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig
export PKG_CONFIG_PATH

6. Now execute the command

sudo ldconfig -v

Making your json file config

Connect your realsense camera and run

realsense-viewer

Then make any setting changes you wish to make and save changes using the save icon on the camera. You will be prompted to save it to a .json file We will denote this json file as YourJsonFile.json

Save this file in the Code/ directory

Directions for Use

Then plug in your Intel RealSense Camera into the computer's USB port

Then run:

cd ~/DepthCameraRecordingTool/Code/ && mkdir Testing_DIR

Then run:

make -j4

Or for a new run try

make clean && make -j4

Then in order to see a list of options do the following

./bin/multicam -h

Examples

For example here is a simple recording for 30 seconds with files saved to Testing_DIR/

Decrease crf for better quality at expense of higher cpu processing and possible frame delays thr just means number of cores you wish to utilize, default is 2, choose a number greater than 8 or 0 to cause it to use all cores

./bin/multicam -dir Testing_DIR/ -sec 30 -fps 30 -crf 23 -thr 0 -jsonfile Default.json

Record with depth, color and IR stream while aligning depth frames to color

./bin/multicam -dir Testing_DIR/ -sec 30 -fps 30 -crf 23 -thr 0 -jsonfile Default.json -color 1 -ir 1 -align_to_color 1

Record with just depth, and IR stream for 30 seconds

./bin/multicam -dir Testing_DIR/ -sec 30 -fps 30 -crf 23 -thr 0 -jsonfile Default.json -ir 1

For command info help type:

./bin/multicam -h

Run multicam program

make -j4 && ./bin/multicam -sec 60 -thr 4 -fps 30 -crf 22 -numraw 450 -max_depth 1000 -dir Testing_DIR/ -bagfile ~/Downloads/fishy-fish.bag && ./bin/decompress Testing_DIR/test_lsb.mp4 Testing_DIR/test_msb.mp4 

Decompress Lossy

Decompressing and comparing against raw files

./bin/decompress -ilsb Testing_DIR/test_lsb.mp4  -imsb Testing_DIR/test_msb.mp4 -hd Testing_DIR/video_head_file.txt -cmp Testing_DIR -o Testing_DIR -print_psnr 0

Decompress Lossy

Decompressing regular no comparison

./bin/decompress -ilsb Testing_DIR/test_lsb.mp4  -imsb Testing_DIR/test_msb.mp4 -hd Testing_DIR/video_head_file.txt -o Testing_DIR -print_psnr 0

Decompress each raw file seperately

ffmpeg -i Testing_DIR/test_lsb_5000_out.mp4 -i Testing_DIR/test_msb_5001_out.mp4 -map 0:v -c:v rawvideo -pix_fmt yuv420p Testing_DIR/output_lsb_5000.mkv -map 1:v -c:v rawvideo -pix_fmt yuv420p Testing_DIR/output_msb_5001.mkv

Directions for Configuring Server Client System

About

A server-client system for Depth Cameras.

Server-Client Configuration Diagram

# Hardware Requirements. Server computer should have more CPU cores for every client (Recommend for 3 or more systems should be at least 6th gen i5 desktop CPU or equivilant)

Client computers can be found for cheap if used (around $150 or less) for example (a used HP EliteDesk 800 G2 Mini Business Desktop with an i5-6500T is sufficiently powerful enough.)

Our configuration: Server (1): HP EliteDesk 800 G2 Mini Business Desktop with 16GB RAM and 5TB external drive

Clients (multiple):
    HP EliteDesk 800 G2 Mini Business Desktop with 8GB RAM, and 128GB SSD     

Depth Cameras: Intel® RealSense™ D435 Stereo Depth Cameras

Ubuntu 20.04 LTS is installed on each computer. See more configuration settings in the README.md document

First please configure for static IP addresses:

On server

Go to Settings -> Network -> Wired -> [Click on Settings Symbol] -> IPv4 -> Manual:

Then enter the following into their respective boxes
In Addresses section
    Address: 192.168.1.234
    Netmask: 255.255.255.0
    Gateway: 192.168.1.1
In DNS section:
    8.8.4.4,8.8.8.8
Then hit apply

**NOTE IF YOU HAVE ANY FIREWALL SETTINGS **

1. Open a terminal

2. Type in

sudo ufw status verbose

3. You will see either an active status or inactive status

4. If you have an active status add rules so you can accept your client computer cameras

5. Type the following into a terminal

sudo ufw allow proto tcp 192.168.1.201 to any port 5000 && sudo ufw allow proto tcp 192.168.1.201 to any port 5001 && \
sudo ufw allow proto tcp 192.168.1.202 to any port 5002 && sudo ufw allow proto tcp 192.168.1.202 to any port 5003 && \
sudo ufw allow proto tcp 192.168.1.203 to any port 5004 && sudo ufw allow proto tcp 192.168.1.202 to any port 5004 && \
.
.
.
...

If need be you can remove all the firewall rules by doing

sudo ufw status verbose

Looking at the number in the list at which the rule occurs at

sudo ufw delete [x] # Where x is the number in the list at which the rule occurs at which you want to delete

Key generation

Type the following in to a terminal on your Server

cd ~/DepthCameraRecordingTool/Code/ && \
make clean && make -j4 && \
python3 make_ssl_keys_cert.py

On Client computers

Go to Settings -> Network -> Wired -> [Click on Settings Symbol] -> IPv4 -> Manual: Then enter the following into their respective boxes In Addresses section Address: 192.168.1.2xx (Here xx means any two digit number as long as it is unique and not 34) Netmask: 255.255.255.0 Gateway: 192.168.1.1 In DNS section: 8.8.4.4,8.8.8.8 Then hit apply

Copy Entire directory from Server to each client

Copy entire ~/DepthCameraRecordingTool/ directory to storage medium of your choice. Then For each client computer: Navigate on the client computer to your home directory using the Files application on Ubuntu Paste a copy of the DepthCameraRecordingTool in this folder.

Alternativley you could just open up an ssh session to each client and copy the directories over. (This however assumes you are connecting the computers to routers)

For each Client

cd ~/DepthCameraRecordingTool/Code/ && \
make clean && \
make -j4

crontab -e

Then if prompted type in 1 and hit ENTER

Using the arrow keys go down to the very bottom of the file Then add the following text replacing USER with your client username found if you go to settings and users:

@reboot sleep 10; export XAUTHORITY=/home/<USER>/.Xauthority; cd /home/<USER>/DepthCameraRecordingTool/Code/ && python3 client_side.py &

Then hit the keys CTRL+X then Y then ENTER

Then create a new bash script:

sudo nano /etc/NetworkManager/dispatcher.d/70-wifi-wired-exclusive.sh

Then copy and paste (Paste using the shortcut CTRL+SHIFT+V) the following into 70-wifi-wired-exclusive.sh:

#!/bin/bash
export LC_ALL=C

enable_disable_wifi ()
{
    result=$(nmcli dev | grep "ethernet" | grep -w "connected")
    if [ -n "$result" ]; then
        nmcli radio wifi off
    else
        nmcli radio wifi on
    fi
}

if [ "$2" = "up" ]; then
    enable_disable_wifi
fi

if [ "$2" = "down" ]; then
    enable_disable_wifi
fi

Then hit the keys CTRL+X then Y then ENTER

Then run the following command:

sudo chmod a+rx /etc/NetworkManager/dispatcher.d/70-wifi-wired-exclusive.sh

Then reboot the system

sudo reboot

Repeat for each client

Running Server from CLI

Go to Settings -> Network -> Wired -> [Click on Settings Symbol] -> IPv4 -> Manual: Then enter the following into their respective boxes In Addresses section Address: 192.168.1.234 Netmask: 255.255.255.0 Gateway: 192.168.1.1 In DNS section: 8.8.4.4,8.8.8.8 Then hit apply

Running a recording for 30 seconds at a crf of 22, with 4 clients connected

python3 server_side.py --dir Testing_DIR/ --num_clients 4 --json Default.json --time_run 30 --crf 22

Running Server from GUI

Run the following command in mousePad/Code directory

python3 server_gui.py 

You should then see an image like this:

CRF values control the quality of the video. The higher the CRF value the lower the quality. Please keep it between 18-28.

Valid video dimensions are 640x480, and 1280x720

Please Choose a valid JSON to utilize.

You could also enter in your own json file as well for different settings just be sure that these files are also on the client computers.

Directions Collecting_Training_Code/

Running recording lossless compression with GUI

cd ~/DepthCameraRecordingTool/Code/Collect_Training_Code/
python3 collect_aligned_depth_training_vids.py

Decompressing each frame to a TIF file GUI

python3 decompress_training_vids_gui.py

Run collect_raw for 30 seconds using ffv1, with aligned frames from directory Collect_Training_Code/

make -j4 && ./bin/collect_raw -dir Testing_DIR/ -sec 30 -align 1 -jsonfile ../Code/Default.json

Other Helpful Commands for Developers

To look at frame rates

ffprobe Testing_DIR/test_lossy.h265 -count_frames -show_entries stream=nb_read_frames,avg_frame_rate,r_frame_rate

To look at stats of a video please see the following from (https://ffmpeg.org/ffprobe-all.html)

ffprobe -f lavfi movie=Testing_DIR/test_lsb.mp4,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN,lavfi.signalstats.YAVG,lavfi.signalstats.YBITDEPTH,lavfi.signalstats.YDIF

Automate GDB

for i in {1..20}; do gdb -q -ex 'set pagination off' -ex 'set args -dir Testing_DIR/ -sec 15 -numraw 450 -crf 17 -bagfile ~/Downloads/One_Mouse_5_minutes_depth.bag' -ex run  ./bin/multicam -ex quit; done

Play videos side by side ffplay

ffplay -f lavfi "movie=left.mp4,scale=iw/2:ih[v0];movie=right.mp4,scale=iw/2:ih[v1];[v0][v1]hstack"

See info ffmpeg

ffmpeg -h encoder=libx264

Utility Functions found in ~/DepthCameraRecordingTool/Code/util_functions

python3 decompress_from_bag_file.py -f test.bag -o Testing_DIR/

From depth coordinates (feeding in depth frame as png or tif) to xyz world coordinates given a set of points on the image plane in this example I am using 3 coordinates, you can use any number of coordinates,

python3 from_depth_to_xyz.py -cal ../Testing_DIR/intrinsics.json -d Testing_DIR2/depth/depth_frm_40.tif -l "(13,20),(13,30),(640, 360)"

python3 script_sort_depth.py -f /media/<username>/DRIVE/images -f2 /media/<username>/DRIVE/out\(Gouda\)/