by Julián Arcila
2008 left, economically speaking, many lessons and indisputably the certainty that 2009 should be a period of savings and optimization of investment. The business of electronic security cannot be absent from this debate and in that sense it is possible that the year that has just begun will bring with it an avalanche of opportunities for developers of technologies such as intelligent video and video analytics, which, among other features, allow to obtain additional benefits from a video signal, such as image analysis, detection of specific movements or patterns to analysis of various trends, in the case of commercial establishments.
These technologies, among many benefits, allow user companies to obtain significant savings, both when it comes to protecting merchandise and when monitoring large spaces, as it optimizes the disposition of human resources by providing information that would not be available based solely on the human eye.
To make this article in a fairly descriptive way, which allows distributors, integrators and end users to become familiar with the technology, SECURITY SALES invited eight companies to participate: Sony, Verint, Infinova, Pelco, Bosch, Vicon, Lumenera and American Dynamics. Each one provided valuable information to understand in detail what the concept of video analytics implies, this without neglecting that of intelligent video.
To begin with, it is very important to note that this technology should not be confused with Video Motion Detection (VMD), a modality that has been on the market for more than 20 years; according to Wikipedia, the VMD uses simple rules and assumes that every pixel change in the scene is important. The limitation of the VMD - explains the portal - is that it brings with it a lot of false alarms so many users use it simply to reduce storage space.
Claudia Dunphy, from the company Lumenera, points out, in relation to this topic, that video analysis is more selective and that it distinguishes between people, vehicles and objects based on shape, behavior and size; "Alerts are issued with great precision. For example, when a person enters a certain area and stays for a specific time. The percentage of false alarms is low because the system can clearly identify a person, "says this professional.
But to begin by defining, a successful answer was provided by Verint's sales team for Latin America, a company for which intelligent video is a solution that ensures a protected environment, both for customers and for their security corps. "Safety equipment is often lightly spread over vast areas, making rapid response difficult during emergencies. Video analytics applications take the most critical behavioral and interest data from huge amounts of video information collected, allowing security teams to focus on the most determining situations."
For his part, Luis Olivares, an expert engineer in video analysis at Vicon, brings a new element: speed. He explains that with video analysis it is achieved in a fast way what with the human brain takes a series of time that is to analyze an image; in Vicon they use computer algorithms capable of processing huge amounts of digital data captured by video cameras, with an almost unlimited power of processing such data so that the system "can see" and give an answer of what it is seeing.
Finally, Warren Brown, product manager of video intelligence software at Tyco Int. makes a very important differentiation between the concepts of video analysis and intelligent video. First, he explains that in the analytical video the algorithms that analyze the video generate data or certain events; these "analytics" serve specific purposes such as analyzing the numbers on vehicle license plates or licenses or to detect the movement of, for example, a bag. Intelligent video, Walter says, is the set of related and integrated analytics in such a way that they allow the improvement of the workflow. "For example, a set of analytics such as the IntelliVid product, help the investigator in a store to identify when a theft of merchandise occurred, facilitates a quick investigation, the documentation of the case and allows to share the results almost immediately," explains the director of the firm, who provided the answers through Walter Hawrys, President of International W Inc.
An offer that grows
At present, this sector is developing to the point that the offer of solutions is growing. In line with the previous statement, Julio Lizarazo of Sony explains that the availability of solutions in this field ranges from the most basic detection of movements, to the recognition of faces or car registrations. "There are solutions that analyze the temperature which is used in solutions for night vision at long distances; another very interesting solution is the one developed in England, which uses the information from the images to analyze the statistics of football matches, as well as the strategies used by the opponent, the speeds and distances traveled by each of the players. "
He adds that another very common solution today, especially in Latin America, is that of P.O.S. (point of sales), which ensures that what is billed in a supermarket matches what the person is carrying. A camera connected to each recording machine is required.
But Carlos Leschhorn, vice president of sales for Latin America of Infinova, adds a very important fact for the region, in the sense that he affirms that in the present analytical solutions for analog systems can be offered, something that a few years ago was not possible; it also reports a notable improvement in the accuracy of the reports. He says that "nowadays we can reach levels of 75%~85% accuracy in information, which is very effective especially when we talk about facial recognition. Today, analytics can be used in various applications such as medicine, forensics and flow control applications."
Now, if it is a question of giving an idea of what can be done with these applications, Daniel Pereira, systems sales manager of Crockett, a firm that represents Pelco, points out that the latter manufacturer has applications that allow alarms to be issued in situations as variable as the presence of abandoned objects, adaptive motion detection, directional motion detection, object counting and camera sabotage.
In line with the above, Luis Yañez, support engineer at Bosch security systems in Mexico, points out that the uses for video analysis systems are very wide but that not all are tested and available to firms in this area. In this sense, he explains that the most common could be "motion detection by conditions of state, size, color, speed and appearance for use in plate reading. Of course there is facial recognition, but it is only available to a few due to the high hardware and software costs necessary to perform these functions, and they are still not accurate enough in their operation."
More and more advantages
If you ask technology developers or large firms dedicated to producing inventions related to technology, the approach that arises is that they really arise as a response to needs and that what is sought is to simplify the lives of those who use it. In this sense, it seems that video analysis is fulfilling both ideas and is solving three of the biggest concerns of security officials: false alarms, optimization of the workforce and agility in the processing and handling of information.
Julio Lizarazo agrees with the above, who affirms that the main advantage that can be obtained with video analysis is the reduction of false alarms, since through algorithms it has been possible to prevent them from being triggered by the movement of animals or natural elements such as water. Additionally, he comments that these systems monitor, leading the security official to follow up on the events found by the camera. Finally, he notes that intelligent search makes the selection of events or images more efficient and faster and for this case he cites the example that a software can search for the exact moment when a car disappears from a specific location.
Daniel Pereira brings additional data to the concept of optimization and points out that these applications are very useful when you have to handle a large number of cameras. "Video analytics allows the operator to focus on the areas where something is most likely to be happening, maximizing time and resources and increasing the likelihood of detecting incidents before their severity escalates."
For his part, Luis Yañez de Bosch brings up another aspect that although it is related to the above, also touches on the financial part of the system and is the reduction of operating costs. He says that using these systems allows you to move from a reactive system to a proactive one, improving the responsiveness of the guard in charge of monitoring certain areas under predetermined conditions.
It cannot be omitted, of course, that with the use of these systems comes leveraged an improvement in the detection capacity, because according to Warren Brown, several studies have determined that a human being loses the ability to identify suspicious situations after having spent between 20 and 60 minutes watching videos from several security cameras. "The focus of most analytics on the market is to help the researcher and have the system, not the human, monitor the video, pointing to the researcher when something suspicious happens," he says.
The required elements
Basically, a video analysis system consists of three fundamental elements: capture equipment (cameras), processing tools (software) and storage space (video recorder/PC). The question that may arise in Latin America is how the costs of these systems behave compared to conventional ones. The reality is that initially the equipment can be more expensive, but over time they justify the investment by presenting a better use of the images captured and by optimizing aspects such as the cost of maintenance, which can be done remotely.
However, it is essential to understand that all the elements have the same importance, because if the cameras are defective the software will not be able to process the image; a software with low processing capacity can miss the advantages of having state-of-the-art cameras and an unresilhed storage system because it may not have the capacity to host the required data.
Faced with the above, Carlos Lescchorn reaffirms it by saying that in many cases the primary analytical part resides in the camera that captures the image, but without good software that subtracts that information from the camera it will not help us much. On the other hand, the computer where this software is installed plays an important role in the performance of the application. It is important that the computer goes beyond the basic specifications of the manufacturer so that the performance of the application is guaranteed.
But there is another important element that cannot be left aside and this is the architectures. Claudia Dunphy recalls that in the past video analysis systems relied on a limited ability to analyze a few video channels on a DVR or central computer. He adds that the preferred way to implement such a solution today is to use a system with open platform video management software. This type of application balances the balance more towards the software side than the hardware, as it is more flexible and provides more expansion options in the future.
A very important element to finish is that the technology is already available for any place in the world that wants to implement high-quality video surveillance solutions; Although video analysis projects are not yet widespread in Latin America, the truth is that large-scale projects have already been carried out that give a sign of the region's progress in technological matters. The important thing before selling such a solution is to consult with at least three experts from recognized firms and not buy based on price. This is the worst decision.
Leave your comment