The article discusses the problem of CCTV design, which consists in the constant change of recommended criteria for equipment selection caused by the continuous and rapid evolution of technologies. The results of the analysis of existing standards are presented and discrepancies are identified. This trend leads to the destabilization of clear quality criteria and the risk of unreasonable CCTV design. The paper highlights the methods and results of research into changes in the spatial resolution of images from a range of technical characteristics of video cameras based on the relevant theory, which is implemented analytically and confirmed by computer modeling using specialized software. The results of the study, obtained by calculation and modeling, indicated a general discrepancy in the data obtained, which does not exceed the permissible limits. The data presented in the study results will help designers make an informed choice of video cameras, which will reduce the risk of excessive design.
The current problem with CCTV design is the lack of stable, universal criteria for selecting equipment, caused by the continuous and rapid evolution of technology. This was particularly noticeable during the transition from analog to digital systems. This dynamic creates the following dilemmas: destabilization of quality criteria (throughout the history of these systems, quality assessment criteria have constantly changed) and the risk of unjustified overengineering (the appropriate choice of video camera resolution). (O. Kaidyk);
Designers face these questions when selecting equipment for a new system in accordance with modern requirements, which need to be addressed.
The purpose of this work is to establish the influence of technical parameters of video cameras on the image quality of an object when solving existing types of operational CCTV tasks in accordance with the recommendations of existing standards. This will allow for a reasonable approach to the selection of video cameras and reduce the risk of overdesign. The question of determining the limits of the use of these criteria is also of interest.
The results of this analysis can be applied in scientific research and practical tasks related to CCTV design.
Historically, the first CCTV systems, which appeared in the 1950s, were analog in terms of image transmission and were not subject to industry quality standards. The process of selecting video cameras was largely reduced to a simple comparison of technical characteristics and specifications offered by manufacturers as key selling points. This practice led CCTV developers to make a critical mistake – they did not take into account operational requirements, i.e., the real goals and tasks that the system had to perform.
The level of image detail is one of the key factors when choosing a CCTV camera. This indicator, which reflects the clarity of the object, directly depends on the operational tasks that the system has to perform. It was the need to formalize these tasks that led to the emergence of the concepts of “operational task” and “operational standards.” This concept, formulated in the UK in the late 1980s, formed the basis of one of the first global security standards – BS 8418:1987. These standards are of a recommendatory nature.
The BS 8418:1987 standard specified that in the era of analog CCTV, the main criterion for selecting a camera for specific operational needs was the lens. Since the main purpose was to monitor behavior, a test mannequin served as the benchmark for configuring these cameras.
This standard established the first official criteria for CCTV required to perform specific operational tasks. These criteria were based on detail categories determined by the percentage of the frame height occupied by a fulllength image of a person (Table 1).
Due to improvements in analog CCTV resolution, BS 8418:1987 has been replaced by BS 8418:2009. The new edition of the standard has significantly expanded the list of operational tasks to include monitoring, detection, surveillance, recognition, identification, and inspection.
The monitoring task in this classification is basic and involves only assessing the general situation at the facility (e.g., the situation at the entrance or in the lobby). It allows you to identify general changes (e.g., crowds of people) but does not provide the detail needed to distinguish faces and license plate numbers.
The task of detection involves the guaranteed identification of subjects or objects of observation in a controlled area. Observation consists of determining the rough characteristics of the object or subject of observation, followed by monitoring their movements in order to identify unauthorized or potentially dangerous actions.
The goal of recognition is to determine whether a subject or object of observation, or elements thereof, belong to a specific group of objects of observation.
Solving the identification task requires the ability to distinguish fairly small, characteristic details, such as the face of a person entering a facility or the license plate number of a vehicle. Thus, identification is the process of recognizing a subject or object by its inherent or assigned identifying features.
Since, in analog video surveillance systems, resolution is determined by the number of television lines (TVL), the requirements for solving the task set in BS 8418 2009 and 2015 for security systems remained based on what part of the frame vertically should be occupied by the object of surveillance (Table 2). No
Operational task 1 2 3 4 5 6
Monitoring
Tasks and opportunities
controlling the
crowd Guaranteed detection of a person in the frame
Determining a person’s distinctive features, such as
clothing
Recognition of people familiar to
the operator Quality sufficient for human identification 100 % identification capability, which eliminates doubt 5 10 25 50 100 400 5 10 30 60 120 400
In turn, Irish I.S. 199:1987, Australian and New Zealand AS 4806.22006, and other CCTV standards of that time also used similar criteria to solve various operational tasks.
In 2013, the European Committee for Electrotechnical Standardization (CENELEC), seeking to standardize the design of CCTV systems in the European Union, issued standard EN 501327 “Video surveillance systems for security television”.
The EN 501327 standard defined the criteria for selecting cameras and lenses, methods for assessing scenes and lighting, as well as requirements for the number and placement of cameras. At the same time, the criteria for solving operational tasks (detection, recognition, identification) fully corresponded to the categories established in the British standard BS 8418:2009.
The main difference between the European standard was the introduction of an alternative key parameter – “pixel density” per unit width of the object being observed. The introduction of this parameter was prompted by the emergence of new types of cameras on the market that functioned differently from traditional analog cameras. The relevance of these pixel densitybased criteria increased with the spread of megapixel digital CCTV systems and remained throughout the transition period when digital systems coexisted with analog ones, until they began to dominate.
With the dominance of IP and HD cameras on the market, the old methods of determining operational requirements based on the percentage ratio of a person’s height to the frame have become obsolete. Modern digital systems have moved from measuring resolution in television lines (TVL) to pixels. Requirements are now specified as the number of image pixels per 1 meter of the object at a given viewing distance.
Thus, the criteria that existed in global standards for analog systems were gradually improved until CCTV completely transitioned to new digital systems, where the key metric became the “pixel” rather than the “TV line”.
Until recently, the operational tasks assigned to CCTV were described by relevant international standards (Table 3). The requirements for monitoring and detection are the most unified among these standards, while the criteria for performing more complex operational tasks – identification and recognition – vary significantly. No 1 2 3 4 5 6
Changes in the global political situation, a significant increase in terrorist activity, and the
development of artificial intelligence technologies have led to a review of operational tasks and,
accordingly, the emergence of IEC/EN 626764: 2024 [
The characteristic features of the IEC/EN 626764: 2024 operational standard are presented in Table 4.
The data presented makes it clear that the types of operational tasks have changed and, accordingly, the recommendations for CCTV design. No
Characterizes the type of person, gait and behavior, as well as the type and category of vehicle Confirm familiar faces, actions can be tracked, vehicle license plates can be
recognized Identification of people, recognition of vehicles by
model and year of manufacture, license plates are clearly legible 20 40 80 125 250 500 1500
The Overview task should enable the operator to identify an object that has just appeared among other elements of the image on the monitor. Based on the definition of the Overview concept, this task is solved taking into account the actual lighting conditions and is evaluated by the degree of isolation of the control object from the general background of the scene.
The influence of technical parameters of video cameras on the image quality of an object was
studied by L. Wastupranata [
Modeling an image of a group of people [
The author notes that there is a limit to the practicality of choosing camera resolution for each size of object on the screen. Thus, using a 4 MP camera does not offer any advantages in terms of “detection” compared to a 2 MP camera.
A study of the impact of camera resolution on subject recognition showed that even at the same focal length (the size of a person on the screen does not change visually), the quality and detail of images vary significantly depending on the resolution. In particular, a camera with a resolution of 2 megapixels provides sufficient image quality for successful detailed “recognition” of a subject, while images from a 4megapixel camera can be used to solve the task of “identification” (establishing identity). Thus, if the subject occupies the same space on the screen, increasing the camera’s resolution increases the level of the operational task that can be solved.
When solving Perceive tasks, you first need to understand the conditions in which the objects of observation are located. If the position of the objects is static they are immobile this is one approach, and to solve such a task, you need to determine the focal length of the lens and the resolution of the camera. If the objects are in motion, the number of parameters required increases. To solve such tasks, you also need to determine the exposure time of the electronic shutter. The exposure time of the electronic shutter must be set in order to obtain a clear, nonblurred image of the moving object in each frame.
The issue of transitioning from traditional image quality measurement in television lines (TVL)
to pixels, as well as the concept of changing spatial resolution density, have been highlighted and
analyzed in scientific research, in particular in works [
The theory of spatial resolution density (or pixel density) variation is a key tool for designing modern CCTV systems. Its essence lies in determining the distance from the camera to the object at which the number of pixels per 1 meter of the width of the monitored area (linear field of view) reaches a predetermined value.
When solving any operational tasks, it is necessary that the resolution of the monitor screen is not worse than the resolution of the video camera. Changes in spatial resolution values obtained experimentally for cameras with HD image quality on the monitor are illustrated in Figure 1.
The range of distances at which an object must be recognized is quite extensive and lies between the “identification” and ‘detection’ zones. In this regard, it has been proposed to divide the “recognition” distances into three sections: 1. Recognition – the object is located closer to the edge of the “detection” zone (the object is small on the monitor and has poor detail). 2. Medium recognition – the size and quality of the object’s display allows its main details to be described. 3. High recognition – the object is located closer to the “identification” area (allows for a clear description of the object).
Thus, analysis of information sources has shown that spatial pixel density is used as a modern criterion for solving various types of operational CCTV tasks, and their values vary in existing standards. The “pixel density” parameter of the camera takes into account the characteristics of the matrix, the lens, and the distance to the object of observation.
It has been established that increasing the resolution of video cameras leads to an increase in pixel concentration per 1 meter of the linear field of the observation scene, which allows for better quality display of observation objects at a significantly greater distance from the camera.
Based on this, it can be assumed that further development of technologies and requirements for CCTV systems may lead to the emergence of matrices with such an extremely high number of pixels that it will change the current classification of CCTV operational tasks. As a result, another question will arise about the further use of existing recommendations for evaluating video surveillance systems.
An analysis of existing studies [
To solve the problem, it was decided to use the theory of spatial resolution change as a basis, which was implemented analytically and through computer modeling in specialized software IP Video System Design Tool 2025.
If the linear field of view Hpv and the number of pixels in the matrix nw across its width are known, then the number of pixels npw per unit width of this field of view can be determined as: where L – is the distance from the camera to the object being observed, m; hm – is the width of the video camera matrix, mm; f – is the focal length of the lens, mm.
To determine the number of pixels across the width of the matrix, you need to take into account its aspect ratio. If N – is the total number of pixels in the matrix and the aspect ratio is 4:3 (width to height), then the number of pixels across the width is determined as:
Hence, the change in spatial resolution can be defined as:
The focal length, depending on the selected type of task and the distance to the object of observation, is determined as: nn3= n
w = H pv
nw , L hm
f nw=√ 0.75
nm . nsv= √ 0.75
nm L hm f
. f = Lhm nsv .
H pv (1) (2) (3) (4)
All studies were based on modeling changes in spatial resolution for a video camera with a 12 mm lens and a 1/2inch sensor. The following range of parameters was selected for analysis: the resolution of the sensor was 5 megapixels and was increased in increments of 5 megapixels, while the range of distances to the object was from 5 to 100 meters in increments of 5 meters.
The results of the calculations of spatial resolution changes are summarized in Table 5.
Matrix resolution nm,
Distance to the object L, m 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
5 2581 968 484 322 242 193 161 138 121 107 96 88 80 74 69 64 60 56 53 50 48
10 3651 1369 684 456 342 273 228 195 171 152 136 124 114 105 97 91 85 80 76 72 68
15 4472 1677 838 559 419 335 279 239 209 186 167 152 139 129 119 111 104 98 93 88 83
20 5163 1936 968 645 484 387 322 276 242 215 193 176 161 148 138 129 121 113 107 101 96
25 5773 2165 1082 721 541 433 360 309 270 240 216 196 180 166 154 144 135 127 120 113 108
30 6324 2371 1185 790 592 474 395 338 296 263 237 215 197 182 169 158 148 139 131 124 118
Modeling of changes in spatial pixel density in IP Video System Design Tool 2025, in order to confirm the obtained calculation data, was carried out using the following data: focal length of the video camera lens 12 mm; distance from the video camera to the human figure model from 20 to 60 m with a step of 5 m; video camera matrix form factor 1/2"; matrix format 4:3; video camera resolution 5Mpx, 10Mpx, 19Mpx. Since there is no 20 Mpx video camera in the program database, a video camera with a resolution of 19 Mpx was used as the closest to the required value.
The human figure model was placed at a distance of 20 m from the video surveillance camera and, after recording the spatial pixel density readings, was moved 5 m further. The spatial pixel density values were obtained together with the formed image of the human model. The simulation results in the selected range are presented in Figure 2.
The corresponding graphical dependencies are presented in Figure 3.
Having determined the type of operational task to be solved for a specific video surveillance area, the designer can use graphical dependencies (Fig. 3) to select the required video camera resolution. This will help avoid excessive design.
Comparative analysis of data obtained by calculation and modeling revealed a discrepancy in values at a certain distance from the camera from 5 to 2 pix/m. In general, this indicates that the total deviation of the obtained results is no more than 3.
A person, given his physiological capabilities, can recognize a familiar person at a limited distance. This distance is mainly 3035 m. Thus, the spatial density values recommended in the existing CCTV standards for a certain type of operational task can be fully provided at different distances by video cameras available on the market. The presented data of the research results will help designers to reasonably approach the choice of video cameras, which will reduce the risk of overdesign. The use of a 25 Mpx video camera overlaps the capabilities of the human visual apparatus. Therefore, with the free presence of 25 Mpx cameras on the market, it will be possible to move away from the use of the currently applicable criteria for operational tasks.
It is advisable to rely on the recommended criteria for operational tasks of CCTV standard IEC/EN 626764: 2024 when designing video surveillance systems with integrated video analytics and artificial intelligence technologies that perform a key preventive function. These systems require high scene resolution for analysis and are capable of detecting abnormal behavior in real time that may indicate the preparation of a terrorist act. This early detection capability allows law enforcement agencies to gain situational awareness and respond quickly to potential threats before they escalate into a real incident. Thus, in critical sectors, it will be possible to control their perimeter and perform video analytics of intrusions into the area, monitor safety by checking for the absence of helmets, protective eyewear, etc. At transport hubs, it will be possible to ensure passenger safety, control flows, and counter terrorism by detecting abandoned objects and aggressive behavior, analyzing crowds, and detecting queues. The implementation of these criteria in smart city monitoring systems will improve quality of life, public safety, and the efficiency of city services. This applies to the detection of unauthorized dumping of garbage or the start of a fire, improper parking, the detection of outbreaks of panic or mass riots, etc.
Dear organizers of the International Scientific Workshop “Applied Information Technologies and Artificial Intelligence Systems”!
We sincerely thank the organizers for inviting us to participate in the 1st International Scientific Workshop “Applied Information Technologies and Artificial Intelligence Systems.”
We are honored to receive such an invitation. We have reviewed the topics and program of the Workshop with great interest and consider it extremely relevant and important for the development of applied information technologies and artificial intelligence systems.
We are confident that the Workshop will be an excellent platform for exchanging knowledge, experience, and establishing new scientific contacts.
During the preparation of this work, the authors used Gemini and Grammarly in order to: Grammar and spelling check and as a smart Search Engine to find related works based on context of conversation. After using these tools/services, the authors reviewed and edited the content as needed and take full responsibility for the publication’s content.