Surveillance System

Table of Contents

• Overview
• Requirements
• Installation
• Usage
• Options
• Components
• License

Overview

The Surveillance System is a comprehensive application designed for real-time video processing, object detection, event-driven analytics, and cloud-based storage. It operates with three main processes:

xVision: Responsible for frame capture and initial processing.

CVRDaemon: Manages continuous video recording.

SurveillanceSystem: Handles event detection, analytics, and interaction with cloud services.

Features

• Frame Capture: Captures frames from cameras using CaptureFrameFromCamera.

• Object Detection: Utilizes TensorFlow Lite for object detection and processes frames with high efficiency.

• Event-Driven Analytics: Analyzes frames for motion events and other significant activities.

• Video Recording: Records video continuously and saves it locally or uploads it to the cloud.

• Cloud Integration: Uploads significant events and video segments to the cloud for storage and further analysis.

• Efficient Memory Management: Handles frame buffers and TensorFlow Lite interpreter invocation efficiently to minimize memory usage.

Requirements

• CMake 3.5 or higher
• A C++17 compatible compiler
• log4cplus
• curl
• OpenCV (core, imgproc, imgcodecs)
• TensorFlow Lite (if USE_TENSOR_LITE is enabled)
• TVM (if USE_TVM is enabled)
• PlumerAI (if USE_PLUMERAI_MODEL is enabled)

Installation

1. Clone the repository:

Clone the repository:
git clone https://github.com/shivoHum2021/surveillanceSystem.git

2. Create a build directory and navigate into it:

mkdir build
cd build

3. Configure the project with CMake:

cmake ..

4. Build the project:

make

Usage

After building the project, you can run the surveillanceApp executable. The application supports various options and configurations based on the CMake options enabled during the build process.

./surveillanceApp

Options

The project supports several options that can be enabled or disabled at build time:

• USE_TVM: Compile with TVM support (default: OFF)
• USE_TENSOR_LITE: Compile with TensorLite support (default: OFF)
• USE_PLUMERAI_MODEL: Compile with PlumerAI support (default: OFF)
• ENABLE_CLASSIFICATION: Compile with object classification support (default: OFF)

These options can be set when running the CMake configuration step:

cmake -DUSE_TENSOR_LITE=ON -DENABLE_CLASSIFICATION=ON ..

Components

Logger

The logger component is responsible for logging throughout the application. It uses the log4cplus library.

HTTP Client

The HTTP client component handles HTTP requests and responses. It uses the curl library.

Frame Processing

The frame processing component handles operations related to camera frame handling and processing. It uses OpenCV for image processing tasks.

Model Processing

The model processing component provides support for different machine learning frameworks. Based on the options enabled, it can support TensorFlow Lite, TVM, and PlumerAI models.

Main Application

The main application (surveillanceApp) integrates all the components and manages the overall surveillance system operations.

License

This project is licensed under the MIT License. See the LICENSE file for details.