Object detection app

This application uses state of the art algorithms for object detection and a frontend application made with Angular. It can run on Jetson Nano, Raspberry Pi or PC.

It can use:

• SSD Mobilenet
• Yolo
• Motion detection using OpenCV
• Cascade classifier

Install

• Make sure to have git:

sudo apt install git

• Install this object-detection package using the Makefile to have all dependencies installed automatically

git clone https://github.com/cristianpb/object-detection.git
cd object-detection/
make up

Raspberry

For Raspberry Pi install OpenCV 4 fast and optimized for this device:

git clone --single-branch --depth=1 https://github.com/dlime/Faster_OpenCV_4_Raspberry_Pi.git
cd Faster_OpenCV_4_Raspberry_Pi/debs
sudo dpkg -i OpenCV*.deb
sudo ldconfig

For the dependencies, I prefer to use .deb files in Raspberry Pi instead of pip because it doesn't have to compile sources. For installing pandas takes more than 1 hour using pip. More details are in Makefile.

Utilisation

Once the application is running go to localhost:5000

Used detection models

• SSD mobilenet

• Yolo V3