OpenVino scripts for Object Detection

OpenVino code set to run Object Detection models

1. Install Python, Git and GPU drivers (optional)

You may need to install some additional libraries on Ubuntu Linux. These steps work on a clean install of Ubuntu Desktop 20.04, and should also work on Ubuntu 22.>

sudo apt-get update
sudo apt-get upgrade
sudo apt-get install python3-venv build-essential python3-dev git-all

2. Create virtual environment folder

This will be the directory from the root of your home/ directory. We assume you have setup your local git repo in the same directory but this is optional. Change the path to your local git repo.

mkdir /home/<usr>/venv

3. Create a Virtual Environment

python3 -m venv /home/<usr>/venv/openvino_env

4. Activate the Environment

Suggested Local repo on is home//github. If you have it in a differnt location update Path

source /home/<usr>/venv/openvino_env/bin/activate

5. Clone the Repository

Change directory to your local git repo on the device

git clone https://github.com/DataSciNAll/vinoobject.git

6. Install the Packages

This setup installs OpenVINO and dependencies. Install PIP to latest version and install dependencies.

python -m pip install --upgrade pip 
pip install wheel setuptools
pip install -r requirements.txt

7. Setup environment and run python script

Python script is using Public SSDlite_mobilenet_v2 IR file. It has already been converted to IR format. This model is trained from the COCO data set

python3 /home/<usr>/github/vinoobject/object_detection_script.py

OpenCV leverages Video Player to frame the different video segments and it is configured to leverage Source = 0 which should be the webcam define that is streaming from the local machine to the script.