This tutorial guides you through creating a video analytics integration for Network Optix products.

The integration you build will detect and track people, cats, and dogs. To do this, it uses a C++ shared library that integrates the OpenCV DNN module with a pre-trained MobileNet SSD Caffe model for object detection, and the OpenCV Tracking-by-Matching (tbm) module for object tracking.

What you will learn

This step-by-step tutorial covers how to:

• Download the sample source code.

• Set up your development workspace.

• Initialize your project using a template.

• Load and enable the integration inside the VMS Server.

• Implement an object detector and send object metadata to the Server.

• Add object tracking capabilities.

• Implement analytics events.

Prerequisites

Before you begin, ensure you have the following:

• A development environment configured for C++ development.

• Access to the Nx Meta Metadata SDK.

• OpenCV library installed with the DNN and tbm modules.

Important Note on Terminology: 

As we transitioned our platform, we updated our terminology. Please keep the following in mind while reading this documentation:

• Interchangeable Terms: The term Plugin (our legacy terminology) and Integration (our updated terminology) are used interchangeably throughout the text. Where you see one, the other may also apply.

• Strict Code Requirements: > Critical Exception: In all code snippets, variables, and API endpoints, the exact wording provided is strictly required. Substituting "plugin" for "integration" (or vice versa) within the code will cause errors.