In this special guest feature, Anusha Jayasundara, a software engineer at WSO2, takes a look at some common use cases for video analytics, and then examines the underlying technology of this growing field. Anusha is a software engineer on the WSO2 real-time analytics team, where he researches and analyzes various methods of video processing to support solutions, such as surveillance and monitoring.
The global market for video analytics is expected to reach $11.17 billion by 2022, according to a recent research report by MarketsandMarkets. This growth comes as newer analytics technologies are being applied to a range of applications—automated surveillance and traffic control, to name just two—that allow people to gain greater insights from video data and turn them into action.
However, to fully capitalize on video processing, organizations need to leverage the right methodologies for extracting high-level information from video sources such as camera feeds. Let’s first look at some common use cases, and then examine the underlying technology.
With traffic monitoring, video processing can be used to detect vehicles, monitor vehicle density, and track vehicles by reading their license plates. Common applications are to use video feeds from traffic cameras to identify speeding vehicles or to observe the traffic flow speed on highways to, for example, predict travel time or dynamically calculate toll values.
Another use case of video analytics is intrusion detection. Here, the video stream captured from a closed-circuit television (CCTV) camera can be used to track the proximity of an object within a specific area and determine if it’s an intrusion or a threat. Video processing also can be used for surveillance to detect unauthorized access and abnormal behaviors.
Contrary to common belief, the video processing behind these use cases and others is not a single task. Instead, it involves a collection of subtasks that can be broken down into detection, recognition, and tracking of a specific object.
To detect an object, computers can be programmed with specific detection models to differentiate between the various types of objects (e.g. a bird versus a plane) and the models can be used to capture those objects in a video frame. Once a system detects the object, it then maps this object to a known related sample dataset to match the features and recognize unique features, such as faces, number plates, or handwriting. After the object has been recognized, its behavior then needs to be monitored via tracking.
Object tracking can be considered a complex process, since it requires the analysis of subtle elements, such as a person’s poses, lighting conditions, and illuminations. However, several object tracking algorithms and methods exist that can be utilized to do this effectively. Two of particular note are Open CV and Scale Invariant Feature Transform (SIFT). OpenCV is an open source computer vision and machine-learning library specifically designed for image processing and video processing tasks. Meanwhile SIFT is a tracking algorithm that’s almost invariant (i.e. can be relied on to be true) through many changes in an object, like shape, rotation and translation.
The three subtasks comprise the video processing layer of a video analytics architecture, which is used to extract high-level information from camera feeds to meet the requirements of the use cases described earlier. Let’s look further at the layers of this architecture.
First, a real-time streaming protocol is used to send data from video sources to the video processing layer for processing. The sources may include video streaming devices, such as CCTV cameras, traffic cameras, online video feeds, or any other video source.
Next, the processed results are sent to a complex event processing engine where all the input gets filtered and processed to extract the meaningful events. Then a user interface (i.e. dashboard) displays the final results in a readable and understandable manner. These capabilities are available in a number of analytics platforms on the market today that can process data streams in real time.
A mobility and device management layer serves to configure all devices involved throughout the process, from the time that data is collected from devices through to when the video analytics results are displayed on mobile devices.
Finally, an identity and access management layer is added to secure communication throughout the whole process and to manage the system’s security and authentication requirements.
With the right technology and architecture deployment, video analytics can take surveillance and other monitoring tasks to a new level—reducing time, money, and human effort while delivering video-enabled solutions that are more consistent, reliable and secure.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind