German 
English 
PHYTEC France 
PHYTEC USA 
PHYTEC India 
PHYTEC China Total amount:
plus VAT.
The Vision Integration Platform
Complexity clearly separated
The industrial-grade platform for customized embedded vision systems
The Vision Integration Platform, or phyVIP for short, integrates the complete vision foundation onto a production-ready baseboard. Project-specific customization is achieved via a deliberately simple connector board. Scalable AI levels provide additional flexibility for diverse vision requirements. This clear separation of standard and custom components reduces development effort, risk, and time to series production.
Why this separation leads to a faster series
Traditional embedded vision projects generate significant complexity at the board level: power supply, camera interfaces, interfaces, security, and thermal requirements all need to be rethought together. phyVIP shifts this recurring complexity to a fully developed base board. This allows development to focus on the application-specific aspects—shortening development cycles and reducing risk on the path to mass production.
Easily integrate cameras and interfaces
Up to two phyCAM-M board-level cameras can be directly connected to the base board. The drivers are already integrated into the Linux BSP, eliminating the need for additional software integration. Flexible FFC cables allow for free camera positioning, and compatible lenses are also available. All other interfaces are routed via high-density board-to-board connectors. The integrated wide-range power supply from 12 to 36 V DC reduces the need for additional controllers, and special solutions such as PoE can also be implemented via the expansion board. This keeps the platform standardized at its core and flexible in its application.
Three performance levels for your vision application
phyVIP is designed for embedded vision applications where AI is executed directly within the system. The NXP i.MX 8M Plus and NXP i.MX 95 processors are available, depending on the application's computing power, interfaces, and scalability requirements.
Building on this, phyVIP can be configured in three AI expansion stages: Core, Edge, and Max. This creates the right combination of processor platform and AI performance — from the compact vision base to the high-performance platform for demanding workloads.
Frequently Asked Questions
phyVIP is a modular vision integration platform for industrial embedded vision and edge AI applications. The platform combines a production-ready base board with a project-specific connector board. The base board already integrates key functions such as the processor, power supply, camera connection, and industrial interfaces. Customizations are implemented via the connector board. This significantly reduces development effort, hardware complexity, and time-to-market.
phyVIP was specifically developed for demanding embedded vision and AI applications. Typical use cases include industrial image processing, quality control, object recognition, robotics, smart cameras, autonomous systems, and AI-powered automation. Thanks to its scalable architecture, the platform is suitable for both compact edge AI systems and more powerful multi-camera or real-time vision applications.
Yes. phyVIP was designed for edge AI, allowing AI models to run directly on the device – without a permanent cloud connection. This reduces latency and enables real-time decision-making. At the same time, sensitive image and process data remain local within the system, improving data privacy and reliability. Depending on the application, users can choose between different AI performance levels – from an integrated NPU in the processor to additional AI acceleration.
The clear separation between the standardized base board and the project-specific connector board significantly reduces development effort and risk. The base board already contains complex and validated core functions such as processor integration, power supply, high-speed interfaces, and camera connectivity. Changes or customer-specific adaptations primarily affect the connector board. This allows new variants to be implemented more quickly without requiring a complete hardware redesign each time.
Many key hardware functions of a vision system are already developed, tested, and integrated at phyVIP. Teams therefore don't need to revisit topics like power design, camera integration, high-speed routing, or fundamental system architecture. Instead, they can focus directly on the actual application. This reduces development time, design risk, and NRE effort, significantly accelerating the path to a production solution.
