Cover Image

Not for Sale

View/Hide Left Panel
Click for next page ( 40

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 39
39 transit policing was established by 1979, it may simply stem TABLE 12 from the initial view of surveillance as primarily a patron WHEN VIDEO CAMERAS ARE MONITORED traffic control and rail operations tool rather than a law When Monitored No. enforcement tool. It may also simply reflect a continuation 24 Hours a Day, 7 Days a Week 22 of past practices. During Hours of Service Only 7 Another Configuration 7 VIDEO SURVEILLANCE SYSTEM FEATURES Not Viewed 9 Note: Two agencies checked multiple responses. Many of the questions pertaining to the technological fea- tures of an agency's surveillance system did not receive Although some agencies checked multiple responses, replies or received replies that were internally inconsis- making it difficult to fully interpret the responses, the larg- tent. As with all questionnaires, it is difficult to determine est number of agencies indicated that their cameras are under why some questions are answered and others are not. Many constant monitoring and the smallest number indicated that respondents were vague about when surveillance was intro- the cameras are not viewed at all in real time. The six agen- duced on their system. This, in combination with the few cies that reported that cameras were never monitored in real responses to a request to provide the name of the surveil- time stated that the decision was based on the cost involved lance vendor, supports a tentative conclusion that managers in having personnel assigned to this function, although two responsible for daily operation of the surveillance system are also indicated that other unspecified factors played a role in less interested in the technical specifications of their systems their decisions. than in its day-to-day use and reliability. With this caveat, it can only be stated that most agencies rely on digital rather Once an agency decides that cameras will be monitored, it than analog systems, and that more than a third described must decide who will monitor them (Table 13). Although the their surveillance systems as combined or in transition from synthesis did not delve into many areas of personnel decision- analog to digital. making, such as whether labor agreements were a factor, the agencies that reported that cameras were not viewed listed Particularly given the current attention paid to video cost as the major reason for this decision. In most agencies, analytics, most of the existing surveillance systems were rail operations personnel are assigned to monitor system described by respondents as having what today would be operation during all hours train are running and sometimes considered relatively low-tech features (Table 14). even when they are not. Because cameras serve a number of non-law enforcement purposes, having rail operations per- sonnel monitor video systems is consistent with the camera's TABLE 14 overall roles in safe operations of the rail network. On other VIDEO SURVEILLANCE SYSTEM SPECIAL FEATURES systems, where there is a full-service 24/7 police agency that Special Feature No. monitors emergency telephones and responds to incidents on 24-hour Recording 36 the transit system, having those individuals monitor the sur- Auto Emergency Digital 4 veillance network is also consistent with their roles. Transmission Secondary Power Source 18 TABLE 13 Auto-start 8 PERSONNEL MONITORING VIDEO CAMERAS Low Light Resolution 16 Who Monitors No. Police/Security Personnel 10 Rail Operations Personnel 8 VIDEO ANALYTICS Combined 22 As video surveillance has proliferated two new issues have Note: Not all agencies responded. emerged: perception overload and the expanded use of sen- sors in conjunction with cameras or as stand-alone tools to Without more information, for instance, whether union protect vital areas. Both rely on advanced technology that agreements played a role in the decision, whether the initial a number of agencies are introducing into their video sur- purpose of the video system played a role in the decision, veillance networks. Even in agencies that assign personnel or whether existing communications networks were used to monitor images in real time, the rise in interest in video to activate the surveillance system, it is difficult to general- analytics is based on the realization that most surveillance ize as to how a combined network came to be the preferred systems produce far more images than it is possible for view- method. Recalling that the earliest federal report noted that ers to absorb. The use of video analytics ("smart" or "intel-

OCR for page 39
40 ligent" video) attempts to provide a solution. At its most rail to a sporting event. Furthermore, even before such sys- simple, smart video can be defined as video that thinks for tems have become common, the Federal Bureau of Investiga- you. It not only collects data, but is capable of analyzing the tion (FBI) released a warning that terrorists may be one step data; for instance, in addition to merely filming individuals ahead of smart video. A jihad-advocating website reportedly on a crowded platform, smart video would identify and focus suggested that adherents leave suspicious bags around New on persons who act suspiciously and would alert those in the York and Washington, D.C., to desensitize first responders viewing room to turn their attention to the monitor display- by forcing them to respond to suspicious but harmless items ing this particular action. left in public areas (Weiss and Mangan 2010). Scientists who are studying it and vendors who are mar- NJT has used federal DHS funds to install a system that keting it refer to smart video as the next-generation of elec- is programmed to alert those who monitor the video when tronic video systems. Such systems rely on algorithms to a suspicious activity has occurred. In a station, this might profile behavior based on how people usually behave in cer- be a bag left unattended or in a particular location. Along tain environments and then picking out those whose behav- the ROW, it might be a boat docked under a bridge (Hecker ior is different from others or inappropriate for the location 2006). Also in conjunction with DHS, MTA-MD has been or situation. These systems take into consideration changes developing a smart video system in the Baltimore subway in lighting conditions, an important factor for rail facilities system, light rail stations, and in Maryland commuter trains and parking locations, and can track people as they move (Nakanishi 2009, p. 23). St. Louis' MetroLink combines from one camera to the next. tunnel intrusion with analytics to monitor its tracks and tunnel. The intrusion sensors indicate activity in the area Understanding and Using Analytics while the analytics are able to determine whether the intru- sion is authorized or not (Resnick 2009). Boston's MBTA is Video analytic systems analyze data to improve tracking. using smart video elements in its recently updated camera They are programmed based on what people can be antici- network, particularly in and around tunnels. Smart video is pated to do. If normal behavior can be anticipated, abnormal also a large component of the National Capital Region Rail behavior can be made to stand out. For instance, to track an Pilot Program involving Amtrak, as discussed in Chapter individual at an airport, the system is provided with infor- three. The use of analytics is also a feature of the surveil- mation on the routes people are likely to take. The system lance systems of two case study agencies, Metro Transit and understands and absorbs that most people go from the air- Valley Metro (see chapter five). port entrance directly to the ticket check-in area, most likely then to check the flight information board, and from there to Distinguishing the Usual from the Unusual security checkpoints. Because it is designed to detect behav- ior that differs from the norm, the analytic-based system is Because of the greater focus on airport security than rail intended to pick up someone who follows no logical pat- security, a demonstration of smart video in late 2009 at tern through the facility. This could be someone who stops airports in the United Kingdom used footage obtained at and then starts moving again in an erratic pattern. It could Heathrow International Airport, where a group of scien- be someone who seems to linger in front of doors that are tists said their prototype identified potential threats that alarmed or marked "employees only," suggesting that the human operators would have missed (Fleming 2009). In a person might be considering whether it is possible to enter study conducted among Florida transit agencies, Dmitry B. without detection. Goldgof and colleagues (2009) found that few agencies were knowledgeable about analytics. The study also referred to a Although some transit agencies are making use of video number of drawbacks, including an analytic system's vulner- analytics, introducing smart video into the rail environment ability to environmental variables such as detrimental light- presents a number of challenges. The major challenge is ing conditions and weather, both of which may lead to false anticipating patterned behavior. This is more difficult in a alarms that could become a source of frustration for the user. transit facility than at an airport, where the most people are Another drawback, particularly in environments where not boarding or alighting from a plane. At a large urban transit all activities can be anticipated, was that to properly program facility, people may be shopping, walking through the station an analytic system, events need to be predefined; events that to avoid city streets in bad weather, dining at one of the facil- have not been defined will not be detected (p. vi). ity's sit-down or fast-food restaurants, or doing any number of things that do not involve taking a train. This is less likely Announcements on breakthroughs in the area of ana- to occur at a small, suburban light rail station, where virtu- lytic software appear regularly in the security and tech- ally all those on the platform are apt to be waiting for a train, nology trade press, which makes it difficult for operations but patterns may still be different if the Monday-to-Friday managers to keep up with the changing technology. For crowd is primarily commuters carrying only briefcases and instance, in the first 2 weeks of June 2010, researchers the evening or weekend crowd is made up of families taking announced that a computer vision system that was not yet