OMNIVISION Launches Low-Power OAX7700 AI Processor for Vehicle Security Systems
OMNIVISION, a global developer of semiconductor solutions spanning advanced digital imaging, analog technologies and display solutions, has introduced the OAX7700, an ultralow-power application-specific integrated circuit (ASIC) designed to bring always-on artificial intelligence capabilities to automotive camera systems.
The new OAX7700 is designed as an always-on vision (AOV) AI companion processor that can transform existing exterior vehicle camera applications, including surround-view systems (SVS), into intelligent parking surveillance platforms. By combining continuous environmental monitoring with edge-based AI processing, the processor is intended to improve vehicle security while significantly reducing the power required for 24/7 operation.
The introduction comes as automakers increasingly integrate cameras and intelligent sensing technologies into vehicles for applications extending beyond conventional driver assistance. Exterior cameras that were traditionally used for parking assistance, surround-view imaging and maneuvering can now support additional functions such as security monitoring, object detection and event recording.
A major challenge in expanding these applications has been the power required to keep camera systems and computing platforms active when a vehicle is parked. OMNIVISION says the OAX7700 addresses this challenge through an ultralow-power architecture that allows the vehicle to continuously monitor its surroundings while remaining in an energy-efficient state until an event requiring further analysis is detected.
Always-On Monitoring With Reduced Power Consumption
The OAX7700 has been developed specifically for always-on automotive vision applications. In a conventional system, continuously monitoring multiple exterior cameras can place a significant demand on the vehicle’s electrical architecture and central computing resources. This can be particularly important when the engine is off or an electric vehicle is parked, because unnecessary energy consumption can affect battery availability.
OMNIVISION states that parking surveillance systems using the OAX7700 can consume up to 97% less power than existing solutions currently available on the market. The reduction is intended to make continuous vehicle monitoring more practical while minimizing the impact on the vehicle’s battery.
Instead of continuously operating at the same processing level, the OAX7700 is designed to remain in an ultralow-power always-on state. It monitors camera feeds for specific visual events and uses its integrated neural processing unit (NPU) to identify potentially relevant activity.
When a significant event is detected, the system can immediately transition into a normal operating mode. This architecture allows the vehicle to maintain continuous surveillance without requiring the full computing platform to remain active around the clock.
Integrated NPU Enables Edge AI Processing
One of the central features of the OAX7700 is its integrated NPU, which provides dedicated hardware for artificial intelligence and machine-learning workloads.
The NPU supports functions including human presence detection, person and object detection, and distance measurement. These capabilities allow the processor to analyze visual information directly at the edge rather than sending every camera frame to a larger central computing platform.
Edge AI can provide several advantages in automotive applications. Processing information locally can reduce the amount of data that needs to be transferred between vehicle systems, lower the workload placed on centralized computing resources and enable faster responses to detected events.
For vehicle security applications, this can be particularly valuable. A parked vehicle may encounter a wide variety of movements, including pedestrians, animals, falling leaves, moving branches, passing vehicles and other environmental activity. A system that simply responds to any movement could generate frequent false alarms.
The OAX7700 is designed to distinguish relevant objects and activity from non-threatening environmental motion. This enables the vehicle to respond to confirmed events rather than treating every detected movement as a security incident.
Pre-Roll Recording Supports Event Verification
Another important feature of the OAX7700-based architecture is its ability to support pre-roll video recording.
When an event triggers the system, the camera can capture approximately five to 10 seconds of video from before the triggering event. This information can then be analyzed by the vehicle’s electronic control unit (ECU) to determine whether the event represents a genuine security concern.
Pre-roll recording can help improve the accuracy of vehicle security systems because the system does not have to begin recording only after an event has already occurred. Instead, the captured footage can provide context surrounding the incident.
For example, if an object approaches a parked vehicle, the system can use the preceding video to determine what happened immediately before the trigger. This additional context may help the ECU distinguish between a genuine intrusion attempt and harmless activity around the vehicle.
The same recorded footage can also provide useful forensic evidence. If an actual incident occurs while a vehicle is parked, owners could potentially use the recorded material to understand what happened and when the event occurred.
Reducing False Alarms
False alarms remain a significant challenge for vehicle surveillance technologies. Outdoor environments are constantly changing, and camera systems can detect movement generated by weather conditions, vegetation, animals and other objects.
Wind-blown trees, birds, leaves, debris and other moving objects can all create visual activity that may resemble a potential threat to a basic monitoring system.
The OAX7700 uses its edge AI capabilities to help address this issue. Instead of responding to motion alone, the processor can analyze detected objects and their characteristics before triggering a response.
The system is designed to trigger when a confirmed threat object is detected, helping reduce unnecessary alerts generated by non-threatening movements.
This approach can make parking surveillance more useful to vehicle owners because it can reduce the number of irrelevant notifications while maintaining continuous monitoring.
Reusing Existing Automotive Camera Systems
OMNIVISION is positioning the OAX7700 as a companion processor capable of enhancing existing exterior camera solutions.
Many modern vehicles already contain multiple cameras for applications such as surround-view monitoring, parking assistance and other advanced vehicle functions. Rather than requiring an entirely separate camera infrastructure for parking surveillance, the OAX7700 can enable these existing camera systems to support additional intelligent functions.
This approach could help automakers expand vehicle functionality without necessarily adding a completely separate sensing architecture.
Surround-view systems typically combine images from multiple exterior cameras to create a comprehensive view around the vehicle. With additional AI processing, those cameras can potentially become part of a broader vehicle security system capable of monitoring activity while the vehicle is parked.
The OAX7700 therefore reflects a broader trend toward multifunctional automotive hardware, where sensors and computing components can support multiple software-defined applications throughout a vehicle’s lifecycle.
Supporting Software-Defined Vehicles
The processor is also designed with the development of software-defined vehicles (SDVs) in mind.
Modern vehicles increasingly rely on software to determine how hardware systems operate and what functions they provide. As vehicles become more connected and software-centric, automakers are looking for hardware platforms that can accommodate new algorithms and features after a vehicle has been manufactured.
The OAX7700’s NPU allows customers to develop their own AI-based algorithms for applications such as human detection, object classification and distance estimation.
This gives automakers and technology developers flexibility to customize the processor according to their specific requirements. Instead of being restricted to a fixed set of functions, the platform can support evolving AI applications.
OMNIVISION also says developers can deploy sophisticated software updates over the air. This capability is important for software-defined vehicle architectures because it allows vehicle functionality to evolve after deployment.
Over-the-air updates can potentially introduce new algorithms, improve existing detection capabilities and refine system performance without requiring physical changes to the vehicle’s hardware.
Reducing the Load on Central Computing Platforms
Another potential benefit of the OAX7700 is the reduction of processing demand on centralized vehicle computing systems.
As automakers add more cameras and intelligent functions, central computing platforms can face increasingly demanding workloads. Sending every image frame from multiple cameras to a central processor for continuous analysis can require substantial computing resources.
By performing AI processing directly at the camera or companion-processor level, the OAX7700 can filter information before it reaches the central compute platform.
Only relevant events or information requiring additional analysis need to be passed to higher-level systems. This distributed processing model can help improve the overall efficiency of the vehicle’s electronic architecture.
For automakers developing increasingly complex electrical and electronic architectures, such distributed intelligence can become an important part of managing computing requirements across different vehicle functions.
Compact Automotive Semiconductor Package
The OAX7700 is supplied in a compact 5.5 mm by 3.5 mm package. Its 60-pin ball grid array (BGA) package integrates stacked PSRAM and flash memory.
The compact form factor is intended to support integration into automotive camera systems where available space can be limited.
Integrating memory alongside the processor can also help simplify system design and support the processor’s always-on AI functionality.
The combination of dedicated AI processing, integrated memory and a compact package gives system designers a semiconductor platform intended specifically for low-power vision applications.
Key Features of the OAX7700
The OAX7700 includes several capabilities aimed at automotive security and intelligent vision applications:
- Ultralow-power vision AI processing architecture
- Integrated neural processing unit for edge AI workloads
- Support for continuous 24/7 always-on operation
- Human presence and person detection
- Object detection and classification
- Distance estimation
- Support for customer-developed AI algorithms
- Reduced false triggers from non-threatening environmental movement
- Five to 10 seconds of pre-roll video recording capability
- Stacked PSRAM and flash memory
- Compact 60-pin BGA package
- 5.5 mm by 3.5 mm package dimensions
- Support for software and algorithm updates over the air
Together, these capabilities are designed to provide vehicle manufacturers with a platform that can combine low-power operation with intelligent visual analysis.
Expanding the Role of Automotive Cameras
The OAX7700 highlights how automotive cameras are increasingly evolving from simple imaging devices into intelligent sensing platforms.
Historically, exterior cameras were primarily associated with driver assistance and parking-related functions. However, improvements in AI processing are allowing the same camera infrastructure to support a wider range of applications.
Vehicle security represents one such opportunity. A parked vehicle can potentially use its cameras to monitor the surrounding environment, identify people or objects and respond to suspicious activity without keeping the vehicle’s entire computing architecture fully active.
The ability to perform these tasks with substantially lower power consumption could become increasingly important as vehicles adopt larger battery systems, more electronic functions and more sophisticated always-connected capabilities.
For electric vehicles in particular, energy efficiency during parked operation can be an important consideration. A low-power security architecture can help maintain surveillance while limiting unnecessary energy consumption.
Availability and Production Timeline
OMNIVISION has announced that samples of the OAX7700 are currently available for customers and developers evaluating the technology. Mass production is scheduled to begin in the first quarter of 2027.
The availability of samples gives automotive customers an opportunity to assess the processor and develop applications ahead of volume production.
As vehicle manufacturers continue to adopt software-defined architectures and increase the number of intelligent functions supported by onboard cameras, processors such as the OAX7700 could contribute to the development of new vehicle security and edge AI applications.
OMNIVISION’s latest semiconductor offering demonstrates the company’s focus on combining imaging technology with dedicated AI processing. By bringing neural processing closer to the camera and dramatically reducing the power required for continuous monitoring, the OAX7700 is designed to turn existing automotive camera systems into more capable, intelligent and energy-efficient sensing platforms.
With samples available now and mass production planned for the first quarter of 2027, the OAX7700 represents another step toward always-on automotive vision systems in which cameras can continuously monitor their surroundings, identify meaningful events and provide intelligent responses while minimizing demands on the vehicle’s overall computing and power infrastructure.
Source Link:https://www.businesswire.com/
