HAWK – Industrial image processing & machine vision

With our Ezpire building block for HAWK We bring MVTec's leading standard software for industrial image processing to your PHYTEC Vision System. Whether you want to implement sophisticated deep learning algorithms, precise measurement technology, or classic defect detection – our Vision BSPs provide the ideal foundation. Test its performance free of charge with the integrated HPEEK demo tool or use your existing full HALCON licenses for direct, hardware-level access to our phyCAM modules.

What is Halcon and what is HPEEK?

MVTec HALCON HDevelop is the comprehensive standard software for industrial image processing with an integrated development environment. It is used worldwide in automation technology, robotics, and quality control.

HPEEK is the associated, license-free embedded evaluation software from MVTec. It allows you to test the performance of HALCON on embedded devices – such as our NXP i.MX phyBOARDs – in advance.

In the PHYTEC context, we ensure that HALCON communicates seamlessly with the underlying Linux architecture and the V4L2 interfaces of our cameras, so that you can run your HDevelop scripts directly on our boards.

HALCON out-of-the-box

Setting up extensive software suites on embedded systems can be time-consuming. We simplify this process for you:

  • Pre-made FTP images: In addition to our standard Vision BSP, we offer specially pre-compiled SD card images on our FTP server, in which HALCON/HPEEK and the PHYTEC demos are already pre-installed.

  • Two-way evaluation: Use HPEEK completely license-free for initial performance tests or connect your HALCON license dongle to access the video streams of the phyCAM modules live.

  • Hardware routing made easy: We provide startup scripts that handle the complex connection setup between the camera sensor and HALCON – optionally as a direct stream or pre-processed by a hardware ISP.

  • Display integration: Our scripts are optimized for the Weston/Wayland display server, so that screen outputs and HALCON GUIs are rendered directly and smoothly on connected monitors (e.g. via HDMI).

Technical Deep Dive

Storage management and system architecture
Due to HALCON's extensive functionality and associated deep learning models, the system requires sufficient storage space. When installing manually using our standard BSPs, the SD card's file system partition must be manually expanded to the full card size after flashing (e.g., using the parted tool). Our specially prepared FTP images eliminate this step for you.

Display and user management (Wayland/Weston)
For the graphical output of HALCON scripts, modern NXP processors use the Wayland display manager via the Weston user. Newer Linux distributions may experience permission issues. To avoid these, our setup scripts automatically configure the HALCON directory for the Weston user if necessary. When scripts are executed, they do so in the correct user environment with the display environment variable set (DISPLAY=:0.0), guaranteeing latency-free image output on the monitor.

Sensor Routing: ISI vs. ISP
Once a full HALCON license (with dongle) is used, the software can access the live data stream from the phyCAM camera modules. We offer various routing paths, which are loaded in the boot environment depending on the SoC's capabilities. Before HALCON can access the image data stream, configuration is performed in the camera and the processor's camera interface using the media-ctl tool. Retrieving and forwarding the image data from the video device is conveniently configured in HALCON using a GStreamer pipeline.

  • Direct path (ISI - Image Sensor Interface): Here, HALCON retrieves the data at the end of the GStreamer pipeline. This approach is particularly suitable for receiving and processing data in a "raw" state.

  • ISP path: For processors with an integrated Image Signal Processor (such as the i.MX 8M Plus), startup scripts are used that first route the data stream through the hardware ISP. HALCON then also retrieves the data using a GStreamer pipeline. However, this results in an already optimally color-corrected and debayered image, which significantly reduces the CPU load and frees up valuable processing time for the actual machine vision algorithms.

Recommended Hardware

To evaluate our System on Modules and phyCAM modules in practice, we offer tailored solutions. Embedded Vision Development Kits These bundles contain everything you need to get started: System on Module (SoM), Baseboard, camera, lens and a downloadable pre-installed Vision BSP with HPEEK and the HALCON examples.

For demanding video applications, we particularly recommend these platforms:

Our benchmark for embedded vision – with dual ISP and dedicated AI unit (NPU) for high-performance multi-camera setups.

The next generation of architecture, equipped with massive AI computing power and the latest MIPI-CSI interfaces for high-resolution streaming.

You can find more Vision setups for other SoCs and form factors on our website. Imaging Kit Overview Page.

The PHYTEC phyCAM portfolio
The HALCON integration unfolds its full potential through our Extensive range of industrial-grade camera modules (phyCAM series)Whether rolling shutter or global shutter, monochrome or color, 1 megapixel to 4k resolutions – our hardware and software drivers are perfectly matched.

Which setup best suits your requirements? Camera modules, for example, can be easily exchanged via the standardized phyCAM interface. Talk to us – We will gladly help you choose the right hardware.

Links & Resources

Due to MVTec's licensing structure, getting started with HALCON is somewhat different than with purely open-source tools. You have two options:

1. The fastest way: Pre-compiled FTP images
For selected kits (such as the phyBOARD-Pollux), we offer pre-configured SD card images on our FTP server, in which the Vision-BSP and HPEE/HALCON are already fully installed. Flash this image, boot the system, and start the HPEE/HALCON demo directly. To test the HALCON demos with live camera data, you will need a license dongle from MVTec.

2. Manual installation via the standard Vision BSP
If you are using our regular phytec-vision image:

  1. Flash the image to a microSD card (min. 4GB) and expand the partition size after the first boot.

  2. Register on the MVTec website for embedded vision and download the Linux-ARM (64-bit) version of HPEEK/HALCON.

  3. Copy the archive to your board, unpack it, and adjust the permissions for the display user.

HPEEK Demos vs. Live Camera Access

  • HPEEK Demos: Can be started immediately and without a license via the command line (e.g. for deep learning evaluations).

  • Live phyCAM access: To run your own HDevelop demos with a real camera image, you will need a valid HALCON license. Copy your license file to the board and plug in the MVTec dongle. Our provided startup scripts will then initialize the data path between the camera and HALCON.