English
Türkçe
IBBVision AI Video Analysis Modules
Integration of Analysis Modules with Deep Learning Surveillance
IBBVision Deep Learning Surveillance is an intelligent system that analyzes footage from security cameras using artificial intelligence. One of the software's most important features is its modular architecture. In other words, the system consists of a core infrastructure and independent analysis modules that can be added to this infrastructure afterwards. This allows each institution to build a system tailored to its own needs.
How Does the Modular System Work?
Deep Learning Surveillance software consists of two main parts:
-
Core Infrastructure (Main Software): This is the primary software required for the system to operate. It captures video streams from cameras, forwards these streams to the analysis modules, and processes the results. The core infrastructure does not perform any analysis on its own; analytical capabilities are gained through the modules added to it.
-
Analysis Modules: These are independent software components, each developed to address a specific security or operational need. For example, suspicious behavior detection, fall detection, license plate recognition, facial recognition, and fire detection are each offered as separate modules.
Unlimited Module Addition Capability
A user who licenses the Deep Learning Surveillance software can add an unlimited number of different AI analysis modules to the system. This means users can continuously expand their systems according to their evolving needs.
Example Usage Scenario
Consider a business owner. Initially, this person purchases the Abandoned Object Detection module for workplace security and adds it to the system. The system begins automatically detecting left bags, packages, or suspicious items.
A few months later, the business owner notices an increase in product theft in the store. They then add the Stolen Object Detection module to the existing system. Now the system can detect both abandoned objects and stolen items.
Later, the business owner learns that employees in the warehouse area are at risk of falling and adds the Fall Detection module as well. The system can now perform three different analyses simultaneously.
Over time, the business grows and new needs arise. The owner continues to enhance the system by gradually adding modules such as Violence Detection, Fire and Smoke Detection, and License Plate Recognition.
Important Note: As seen in this example, users can add as many modules as they wish. There is no upper limit on the number of modules that can be added. Each new module requires a separate license, which the user acquires and then activates on the system.
Number of Simultaneously Running Modules
While there is no limit on the total number of analysis modules that can be added, how many modules can run simultaneously depends on the device's processing capacity.
IBBVision Edge AI Vision Computer devices can run between 1 and 64 real-time analysis modules simultaneously. This range is determined by the device's hardware specifications, processor power, and NPU (Neural Processing Unit) capacity.
Factors Affecting Capacity
| Factor | Description |
|---|---|
| Device Model | Models ranging from EVS 04 (4 channels) to EVS 64 (64 channels) have different processing capacities |
| Number of Cameras | The total number of camera feeds being processed simultaneously determines the processing load |
| Image Resolution | Higher resolution images require more processing power |
| Frame Rate | Frames per second (fps) affect the processing load |
| Analysis Complexity | Some modules, such as facial recognition, require more processing power than simpler motion detection |
Capacity Examples in Different Scenarios
Small Business – EVS 04 (4 Channels)
In this scenario, the business has 4 security cameras. The user wants to run 2 different analysis modules on these cameras: Suspicious Behavior Detection and Fall Detection.
-
4 cameras × 2 modules = 8 analysis streams
-
The device's capacity is sufficient for 4 channels, and the system runs smoothly.
Mid-Sized Institution – EVS 16 (16 Channels)
In this scenario, the institution has 16 security cameras. The user wants to run 5 different analysis modules: Intrusion Detection, Abandoned Object Detection, Facial Recognition, License Plate Recognition, and Density Analysis.
-
16 cameras × 5 modules = 80 analysis streams
-
The EVS 16 device has the capacity to handle this load.
Large-Scale Facility – EVS 64 (64 Channels)
In this scenario, the facility has 64 security cameras. The user wants to run 12 different analysis modules: Suspicious Behavior, Fall Detection, Violence Detection, Fire Detection, Intrusion Detection, Abandoned Object Detection, Stolen Object Detection, Facial Recognition, License Plate Recognition, Density Analysis, Vehicle-Pedestrian Separation, and Wrong Direction Detection.
-
64 cameras × 12 modules = 768 analysis streams
-
The EVS 64, as the highest-capacity model, is designed to operate at this density.
Flexibility Provided by the Modular Structure
1. Adaptation to Changing Needs
An institution's security needs may change over time. New threats may emerge, or operational requirements may shift. For example:
-
A face mask detection module may be needed during a pandemic
-
A violence detection module may be required for student safety in a school
-
A fire and smoke detection module may be added to address fire risks in a warehouse
Thanks to the modular structure, users can always add new analysis modules as needed, without having to replace their existing system or purchase a new one.
2. Phased Investment Opportunity
Institutions do not have to purchase all analysis modules upfront. This is a major advantage, especially for organizations with budget constraints.
-
In the first phase, 2–3 modules are purchased for the most urgent needs
-
As the budget permits or new needs arise, additional modules are added
-
Investment costs are spread over time, balancing cash flow
3. Flexibility for Periodic Needs
Some analysis modules may become more important during certain periods:
-
Holiday seasons: Crowd management modules are used more actively
-
Summer months: Fire detection modules take priority
-
Special events: Violence detection and crowd analysis modules are activated
-
Winter months: Fall detection modules become more critical
Users can deactivate modules they do not currently need to use system resources efficiently, and reactivate them when necessary.
4. Using Different Modules for Different Areas
Different analysis modules can be assigned to each camera. For example:
-
Store entrance: Facial recognition and license plate recognition modules
-
Inside the store: Suspicious behavior and stolen object detection modules
-
Warehouse area: Fall detection and fire detection modules
-
Parking lot: License plate recognition and vehicle tracking modules
This flexibility ensures resources are used as efficiently as possible.
Available Analysis Modules
Deep Learning Surveillance software includes a wide range of analysis modules. Below are some of the available modules:
Person and Object Detection Modules
| Module Name | What It Does |
|---|---|
| Human Detection | Detects all people within the camera's field of view and determines their positions |
| Vehicle Detection | Detects and classifies vehicles such as cars, trucks, motorcycles, and buses |
| Object Detection | Detects specific objects such as bags, backpacks, suitcases, weapons, and knives |
| Face Detection and Recognition | Detects faces, matches them against a database, and performs identity verification |
| License Plate Recognition | Automatically reads, recognizes, and records vehicle license plates |
| Vehicle-Pedestrian Separation | Distinguishes between vehicles and pedestrians in traffic |
| Animal Detection | Detects animals such as cats, dogs, and birds and differentiates them from humans |
| Color and Feature Detection | Detects vehicles of a specific color or people wearing specific clothing |
Behavior and Anomaly Analysis Modules
| Module Name | What It Does |
|---|---|
| Suspicious Behavior Detection | Detects actions such as loitering, sneaking, or repeatedly pacing as if casing for theft |
| Anomaly Detection | Identifies unusual situations that deviate from normal patterns |
| Fall Detection | Automatically detects when a person suddenly falls to the ground |
| Violence Detection | Detects physical altercations such as fighting, pushing, punching, and kicking |
| Running and Fleeing Detection | Detects individuals moving significantly faster than normal walking speed |
| Risky Motion Detection | Detects sudden movements, abrupt direction changes, or erratic motions |
| Crowd Gathering Detection | Detects groups of people gathering in a specific area |
| Proximity Detection | Detects social distancing violations |
Security and Environmental Analysis Modules
| Module Name | What It Does |
|---|---|
| Intrusion Detection | Detects unauthorized entries into or exits from designated restricted areas |
| Restricted Area Access | Detects unauthorized access to sensitive zones such as management offices or server rooms |
| Abandoned Object Detection | Detects objects such as bags, suitcases, or packages that remain stationary for a specified time |
| Stolen Object Detection | Detects when a designated object has been removed from its location |
| Wrong Direction Detection | Detects vehicles or individuals moving in the wrong direction in one-way areas |
| Fire and Smoke Detection | Detects early signs of flames, fire, smoke, and soot through camera footage |
| Perimeter Violation Detection | Detects when a designated line or boundary is crossed |
| Dwell Time Detection | Detects when a person or object remains in an area longer than a set time limit |
| Drone and UAV Detection | Detects unmanned aerial vehicles and drones entering the airspace |
Density and Operational Analysis Modules
| Module Name | What It Does |
|---|---|
| Density Analysis | Measures the concentration of people or vehicles in a given area and counts the number of individuals |
| Crowd Management | Identifies overcrowded areas and alerts when capacity is exceeded |
| PPE Compliance Monitoring | Inspects the use of safety equipment such as helmets, vests, and harnesses in work zones |
| Motion Analysis | Detects abnormal or unexpected movement patterns |
| Entry/Exit Counting | Counts the number of people or vehicles entering or exiting a specific point |
| Parking Occupancy Detection | Detects vacant and occupied parking spaces in parking lots |
| Dwell Time Analysis | Measures the amount of time individuals spend at specific points |
| Thermal Anomaly Detection | Works with thermal cameras to detect elevated body temperatures |
Module Addition Process
Adding a new analysis module to the Deep Learning Surveillance software is straightforward:
-
Needs Assessment: The institution's security or operational needs are identified, and it is decided which analysis modules should be added.
-
Licensing: A license for the required analysis module is obtained from IBBVision. Each module is licensed separately.
-
Activation: The acquired license is activated on the Edge AI Vision Computer device. This is typically done by entering a license key into the system.
-
Configuration: Settings for the newly added analysis module are configured. Parameters such as which cameras it will run on, sensitivity levels, and regions of interest are defined.
-
Execution: The module becomes active and begins real-time analysis.
Frequently Asked Questions
Q: How many different analysis modules can I add to a single Edge AI Vision Computer device?
A: There is no limit on the total number of analysis modules that can be added to the system. Users can add an unlimited number of modules as needed.
Q: Can all the modules I add run simultaneously?
A: Yes, but the number of modules that can run simultaneously depends on the device's processing capacity. IBBVision Edge AI Vision Computer devices can run between 1 and 64 modules at the same time.
Q: Can I add a new analysis module later?
A: Yes. Thanks to the modular structure, new analysis capabilities can be easily added through proper licensing at any time. It is not mandatory to purchase all modules during initial setup.
Q: Can I disable modules I am not using?
A: Yes. Modules that are not actively being used can be disabled to prevent them from consuming system resources unnecessarily. They can be reactivated whenever needed.
Q: Can I run different modules on the same camera?
A: Yes. Different analysis modules can be activated for each camera. For example, one camera may perform only fall detection, while another may perform both suspicious behavior and intrusion detection simultaneously.
Q: Do I need to purchase new hardware to add a module?
A: No. New modules can be added through your existing Edge AI Vision Computer device. Adding new modules only requires a software license; no hardware changes are necessary.
The modular structure of the IBBVision Deep Learning Surveillance software offers institutions unparalleled flexibility and freedom. Users can:
-
Add an unlimited number of analysis modules to the system as needed
-
Add new modules as their needs evolve or increase
-
Disable unused modules
-
Assign different modules to different cameras
-
Control costs through phased investments
All of these modules operate 24/7 with uninterrupted, high-performance functionality, thanks to the powerful processing capacity of Edge AI Vision Computer devices. Depending on the device model, between 1 and 64 analysis modules can run simultaneously.
Thanks to this modular approach, institutions are prepared not only for today's needs but also for future security and operational requirements that may arise. The system grows and evolves alongside the institution.
All AI Image Analysis Modules are developed exclusively by IBBVision. Third-party software or modules are not supported. The operational capacity of modules may vary depending on the hardware specifications and configuration of the Edge AI Vision Computer device.