Orchestration class which reads the input, runs the lane and object detectors, displays the detected output and saves output video stream locally. More...

#include "VideoManager.hpp"

Public Member Functions
	VideoManager ()
	Constructs a new VideoManager object.
	VideoManager (const int32_t &p_inputVideoCamera, const std::string &p_yoloFolderPath, const Globals::ObjectDetectorTypes &p_objectDetectorTypes, const Globals::ObjectDetectorBackEnds &p_objectDetectorBackEnds=Globals::ObjectDetectorBackEnds::CPU, const Globals::ObjectDetectorBlobSizes &p_objectDetectorBlobSizes=Globals::ObjectDetectorBlobSizes::ONE)
	Constructs a new VideoManager object for use with an OpenCV camera input (for API use).
	VideoManager (const std::string &p_inputVideoFilePath, const std::string &p_yoloFolderPath, const Globals::ObjectDetectorTypes &p_objectDetectorTypes, const Globals::ObjectDetectorBackEnds &p_objectDetectorBackEnds=Globals::ObjectDetectorBackEnds::CPU, const Globals::ObjectDetectorBlobSizes &p_objectDetectorBlobSizes=Globals::ObjectDetectorBlobSizes::ONE)
	Constructs a new VideoManager object for use with a video file input (for API use).
	VideoManager (const std::vector< std::string > &p_commandLineArguments)
	Constructs a new VideoManager object (for CLI use).
	~VideoManager ()
	Destructs a VideoManager object.
void	SetProperties (const int32_t &p_inputVideoCamera, const std::string &p_yoloFolderPath, const Globals::ObjectDetectorTypes &p_objectDetectorTypes, const Globals::ObjectDetectorBackEnds &p_objectDetectorBackEnds, const Globals::ObjectDetectorBlobSizes &p_objectDetectorBlobSizes)
	Set the properties of the VideoManager object.
void	SetProperties (const std::string &p_inputVideoFilePath, const std::string &p_yoloFolderPath, const Globals::ObjectDetectorTypes &p_objectDetectorTypes, const Globals::ObjectDetectorBackEnds &p_objectDetectorBackEnds, const Globals::ObjectDetectorBlobSizes &p_objectDetectorBlobSizes)
	Set the properties of the VideoManager object.
void	RunLaneAndObjectDetector ()
	Orchestration function which contains the main program loop.
Performance	GetPerformance ()
	Get the performance object to extract performance-related information.
	VideoManager (const VideoManager &p_videoManager)=delete
	Disable constructing a new VideoManager object using copy constructor.
	VideoManager (const VideoManager &&p_videoManager)=delete
	Disable constructing a new VideoManager object using move constructor.
VideoManager &	operator= (const VideoManager &p_videoManager)=delete
	Disable constructing a new VideoManager object using copy assignment operator.
VideoManager &	operator= (const VideoManager &&p_videoManager)=delete
	Disable constructing a new VideoManager object using move assignment operator.

Private Member Functions
void	ToggleDebugMode ()
	Toggles the debug mode and whether to destroy the debug frames.
void	ToggleSaveOutput ()
	Toggles whether the output video stream is saved locally. If toggled on, it creates a new file in the format YYYY-MM-DD-HH:mm:ss-output.mp4, and if toggled off, it closes the output file.
void	Quit ()
	Destroys all windows and releases all input/output videos.

Private Attributes
cv::VideoCapture	m_inputVideo
	Represents the input video stream.
cv::VideoWriter	m_outputVideo
	Represents the output video stream.
cv::Mat	m_currentFrame
	Represents the current frame being processed.
LaneDetector	m_laneDetector
	The lane detector.
ObjectDetector	m_objectDetector
	The object detector.
Performance	m_performance
	The frame time and current/average frames per second performance calculator.
VideoManagerInformation	m_videoManagerInformation
	The VideoManagerInformation struct containing all video manager-related information.
std::chrono::time_point< std::chrono::high_resolution_clock >	m_saveOutputStartTime
	The time that the output started to be saved locally.

Detailed Description

Orchestration class which reads the input, runs the lane and object detectors, displays the detected output and saves output video stream locally.

Definition at line 28 of file VideoManager.hpp.

Constructor & Destructor Documentation

◆ VideoManager() [1/6]

LaneAndObjectDetection::VideoManager::VideoManager ( )

explicit

Constructs a new VideoManager object.

Warning: SetProperties() should be used to initialise the necessary member variables before calling Run().

Definition at line 24 of file VideoManager.cpp.

◆ VideoManager() [2/6]

LaneAndObjectDetection::VideoManager::VideoManager	(	const int32_t &	p_inputVideoCamera,
		const std::string &	p_yoloFolderPath,
		const Globals::ObjectDetectorTypes &	p_objectDetectorTypes,
		const Globals::ObjectDetectorBackEnds &	p_objectDetectorBackEnds = Globals::ObjectDetectorBackEnds::CPU,
		const Globals::ObjectDetectorBlobSizes &	p_objectDetectorBlobSizes = Globals::ObjectDetectorBlobSizes::ONE )

explicit

Constructs a new VideoManager object for use with an OpenCV camera input (for API use).

Parameters

p_inputVideoCamera	Camera Id of the video capturing device to open using the OpenCV VideoCapture class.
p_yoloFolderPath	The folder containing the .cfg and .weights YOLO files.
p_objectDetectorTypes	The type of object detector to use with an option to disable object detection. The tiny version is more performant at the cost of accuracy.
p_objectDetectorBackEnds	The supported backends for the object detector to run on. In theory, GPU should be significantly more performant. For maximum performance CUDA should be used as the backend, however, this requires building OpenCV with CUDA which is not supported with this projects built-in installation script.
p_objectDetectorBlobSizes	The supported blob sizes for the object detector to run with. The larger the blob size the more performant the detector at the cost of performance.

Definition at line 32 of file VideoManager.cpp.

◆ VideoManager() [3/6]

LaneAndObjectDetection::VideoManager::VideoManager	(	const std::string &	p_inputVideoFilePath,
		const std::string &	p_yoloFolderPath,
		const Globals::ObjectDetectorTypes &	p_objectDetectorTypes,
		const Globals::ObjectDetectorBackEnds &	p_objectDetectorBackEnds = Globals::ObjectDetectorBackEnds::CPU,
		const Globals::ObjectDetectorBlobSizes &	p_objectDetectorBlobSizes = Globals::ObjectDetectorBlobSizes::ONE )

explicit

Constructs a new VideoManager object for use with a video file input (for API use).

Parameters

p_inputVideoFilePath	The file path of the video file to open using the OpenCV VideoCapture class.
p_yoloFolderPath	The folder containing the .cfg and .weights YOLO files.
p_objectDetectorTypes	The type of object detector to use with an option to disable object detection. The tiny version is more performant at the cost of accuracy.
p_objectDetectorBackEnds	The supported backends for the object detector to run on. In theory, GPU should be significantly more performant. For maximum performance CUDA should be used as the backend, however, this requires building OpenCV with CUDA which is not supported with this projects built-in installation script.
p_objectDetectorBlobSizes	The supported blob sizes for the object detector to run with. The larger the blob size the more performant the detector at the cost of performance.

Definition at line 46 of file VideoManager.cpp.

◆ VideoManager() [4/6]

LaneAndObjectDetection::VideoManager::VideoManager ( const std::vector< std::string > & p_commandLineArguments )

explicit

Constructs a new VideoManager object (for CLI use).

Usage: lane-and-object-detection –input ... –yolo-folder-path ... –object-detector-type ... [optional]

OPTIONS:

Generic Options:

-h –help Display available options.

Required Options:

-i –input File path or camera ID -y –yolo-folder-path Path to the yolo folder

Optional options:

-o –object-detector-type One of: none, standard or tiny. (default = none) -b –object-detector-backend One of: cpu or cuda (default = cpu) -s –object-detector-blob-size One of: 208, 320, 416, 512 or 608 (default = 208)

Parameters

p_commandLineArguments List of all command line arguments.

Definition at line 60 of file VideoManager.cpp.

◆ ~VideoManager()

LaneAndObjectDetection::VideoManager::~VideoManager ( )

Destructs a VideoManager object.

Definition at line 170 of file VideoManager.cpp.

◆ VideoManager() [5/6]

LaneAndObjectDetection::VideoManager::VideoManager ( const VideoManager & p_videoManager )

delete

Disable constructing a new VideoManager object using copy constructor.

Parameters

p_videoManager The VideoManager to copy.

◆ VideoManager() [6/6]

LaneAndObjectDetection::VideoManager::VideoManager ( const VideoManager && p_videoManager )

delete

Disable constructing a new VideoManager object using move constructor.

Parameters

p_videoManager The VideoManager to copy.

Member Function Documentation

◆ SetProperties() [1/2]

void LaneAndObjectDetection::VideoManager::SetProperties	(	const int32_t &	p_inputVideoCamera,
		const std::string &	p_yoloFolderPath,
		const Globals::ObjectDetectorTypes &	p_objectDetectorTypes,
		const Globals::ObjectDetectorBackEnds &	p_objectDetectorBackEnds,
		const Globals::ObjectDetectorBlobSizes &	p_objectDetectorBlobSizes )

Set the properties of the VideoManager object.

Parameters

p_inputVideoCamera	Camera ID of the video capturing device to open using the OpenCV VideoCapture class.
p_yoloFolderPath	The folder containing the .cfg and .weights YOLO files.
p_objectDetectorTypes	The type of object detector to use with an option to disable object detection. The tiny version is more performant at the cost of accuracy.
p_objectDetectorBackEnds	The supported backends for the object detector to run on. In theory, GPU should be significantly more performant. For maximum performance CUDA should be used as the backend, however, this requires building OpenCV with CUDA which is not supported with this projects built-in installation script.
p_objectDetectorBlobSizes	The supported blob sizes for the object detector to run with. The larger the blob size the more performant the detector at the cost of performance.

Definition at line 175 of file VideoManager.cpp.

◆ SetProperties() [2/2]

void LaneAndObjectDetection::VideoManager::SetProperties	(	const std::string &	p_inputVideoFilePath,
		const std::string &	p_yoloFolderPath,
		const Globals::ObjectDetectorTypes &	p_objectDetectorTypes,
		const Globals::ObjectDetectorBackEnds &	p_objectDetectorBackEnds,
		const Globals::ObjectDetectorBlobSizes &	p_objectDetectorBlobSizes )

Set the properties of the VideoManager object.

Parameters

p_inputVideoFilePath	The file path of the video file to open using the OpenCV VideoCapture class.
p_yoloFolderPath	The folder containing the .cfg and .weights YOLO files.
p_objectDetectorTypes	The type of object detector to use with an option to disable object detection. The tiny version is more performant at the cost of accuracy.
p_objectDetectorBackEnds	The supported backends for the object detector to run on. In theory, GPU should be significantly more performant. For maximum performance CUDA should be used as the backend, however, this requires building OpenCV with CUDA which is not supported with this projects built-in installation script.
p_objectDetectorBlobSizes	The supported blob sizes for the object detector to run with. The larger the blob size the more performant the detector at the cost of performance.

Definition at line 202 of file VideoManager.cpp.

◆ RunLaneAndObjectDetector()

void LaneAndObjectDetection::VideoManager::RunLaneAndObjectDetector ( )

Orchestration function which contains the main program loop.

Definition at line 229 of file VideoManager.cpp.

◆ GetPerformance()

Performance LaneAndObjectDetection::VideoManager::GetPerformance ( )

Get the performance object to extract performance-related information.

Returns: Performance The performance object containing performance-related information.

Definition at line 283 of file VideoManager.cpp.

◆ operator=() [1/2]

VideoManager & LaneAndObjectDetection::VideoManager::operator= ( const VideoManager & p_videoManager )

delete

Disable constructing a new VideoManager object using copy assignment operator.

Parameters

p_videoManager The VideoManager to copy.

◆ operator=() [2/2]

VideoManager & LaneAndObjectDetection::VideoManager::operator= ( const VideoManager && p_videoManager )

delete

Disable constructing a new VideoManager object using move assignment operator.

Parameters

p_videoManager The VideoManager to copy.

◆ ToggleDebugMode()

void LaneAndObjectDetection::VideoManager::ToggleDebugMode ( )

private

Toggles the debug mode and whether to destroy the debug frames.

Definition at line 288 of file VideoManager.cpp.

◆ ToggleSaveOutput()

void LaneAndObjectDetection::VideoManager::ToggleSaveOutput ( )

private

Toggles whether the output video stream is saved locally. If toggled on, it creates a new file in the format YYYY-MM-DD-HH:mm:ss-output.mp4, and if toggled off, it closes the output file.

Definition at line 301 of file VideoManager.cpp.

◆ Quit()

void LaneAndObjectDetection::VideoManager::Quit ( )

private

Destroys all windows and releases all input/output videos.

Definition at line 335 of file VideoManager.cpp.

Member Data Documentation

◆ m_inputVideo

cv::VideoCapture LaneAndObjectDetection::VideoManager::m_inputVideo

private

Represents the input video stream.

Definition at line 204 of file VideoManager.hpp.

◆ m_outputVideo

cv::VideoWriter LaneAndObjectDetection::VideoManager::m_outputVideo

private

Represents the output video stream.

Definition at line 209 of file VideoManager.hpp.

◆ m_currentFrame

cv::Mat LaneAndObjectDetection::VideoManager::m_currentFrame

private

Represents the current frame being processed.

Definition at line 214 of file VideoManager.hpp.

◆ m_laneDetector

LaneDetector LaneAndObjectDetection::VideoManager::m_laneDetector

private

The lane detector.

Definition at line 219 of file VideoManager.hpp.

◆ m_objectDetector

ObjectDetector LaneAndObjectDetection::VideoManager::m_objectDetector

private

The object detector.

Definition at line 224 of file VideoManager.hpp.

◆ m_performance

Performance LaneAndObjectDetection::VideoManager::m_performance

private

The frame time and current/average frames per second performance calculator.

Definition at line 229 of file VideoManager.hpp.

◆ m_videoManagerInformation

VideoManagerInformation LaneAndObjectDetection::VideoManager::m_videoManagerInformation

private

The VideoManagerInformation struct containing all video manager-related information.

Definition at line 234 of file VideoManager.hpp.

◆ m_saveOutputStartTime

std::chrono::time_point<std::chrono::high_resolution_clock> LaneAndObjectDetection::VideoManager::m_saveOutputStartTime

private

The time that the output started to be saved locally.

Definition at line 239 of file VideoManager.hpp.

The documentation for this class was generated from the following files:

include/helpers/VideoManager.hpp
source/helpers/VideoManager.cpp

Public Member Functions

Private Member Functions

Private Attributes

Detailed Description

Constructor & Destructor Documentation

◆ VideoManager() [1/6]

◆ VideoManager() [2/6]

◆ VideoManager() [3/6]

◆ VideoManager() [4/6]

◆ ~VideoManager()

◆ VideoManager() [5/6]

◆ VideoManager() [6/6]

Member Function Documentation

◆ SetProperties() [1/2]

◆ SetProperties() [2/2]

◆ RunLaneAndObjectDetector()

◆ GetPerformance()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ ToggleDebugMode()

◆ ToggleSaveOutput()

◆ Quit()

Member Data Documentation

◆ m_inputVideo

◆ m_outputVideo

◆ m_currentFrame

◆ m_laneDetector

◆ m_objectDetector

◆ m_performance

◆ m_videoManagerInformation

◆ m_saveOutputStartTime