The standout feature of the XM2 is its integrated , which allows complex deep learning models to run directly on the camera . This on-device processing eliminates the need for external PCs and costly GPUs, reducing system costs and latency.
Perhaps the most impactful announcements from IMAGO in 2021 concerned the "AI" series, a duo designed to make deep learning accessible for both expert developers and absolute beginners.
Because the training and inference processes happen entirely on the camera unit, sensitive product data does not need to leave the production line network, ensuring high data security. The Go to product viewer dialog for this item.
| Feature / Model | | Vision Cam EB | Vision Cam AI | Vision Cam AI.go | | :--- | :--- | :--- | :--- | :--- | | Primary Focus | High-performance, programmable Edge AI computer vision | High-speed, dynamic motion capture | Flexible, powerful deep learning inference | "No-code", ready-to-use deep learning | | Processor | NVIDIA Jetson Orin | ARM Cortex-A15 | Quad-Core ARM + Google Coral | Integrated processor (not specified) | | Ideal User | Advanced developers, system integrators, OEMs | Experts needing high dynamics | Experienced AI developers | End-users, automation technicians | | Sensor | Area scan (5-12 MP) or Line scan | Event-based (Prophesee) | 5 MP CMOS Global Shutter | 5 MP CMOS Global Shutter | | Programming | Fully programmable (C++, Python, OpenCV) | Via API | Fully programmable (C++, Python, HALCON) | No programming, Web GUI | | Typical App | Quality inspection, robot guidance, inline inspection | Vibration analysis, fast counting, welding | Defect detection, code reading | Simple classification, presence verification |
The technical specs of the base VisionCam units deployed during this production cycle emphasize long-term operational stability: Vision Cam XM2: Smart Camera for Edge AI - IMAGO imago visioncam 2021
Its low-light performance and motion detection capabilities enhance security monitoring systems, providing clearer images in a variety of conditions.
: Inspecting complex blister pack seals, verifying pill counts, and recognizing text printing errors where traditional rule-based programming throws high rates of false-positives.
shifted this entire architecture onto the edge. It consolidated the entire pipeline—image acquisition, neural network training, and real-time inference—directly onto the device itself. Code-Free "Plug-and-Play" Training The defining trait of the Vision Cam 2021
: Processing happens entirely on the device, meaning no data is sent to the cloud and no external PC is needed. The standout feature of the XM2 is its
Parallel to the AI-focused models, the VisionCam XM (updated in early 2021) serves as a "personal vision sensor" for developers who
Users can load sample images directly onto the camera and define classes (e.g., "Good" vs. "Bad" or "Type A" vs. "Type B"). The camera then independently trains its neural network, notifying the user upon completion. Data Security:
: Detecting material defects or assembly errors that are too subtle for traditional "rule-based" cameras.
Running on a environment, developers can leverage C++ , Python , OpenCV , and, optionally, over 2,000 HALCON operators . This open architecture provides the freedom to implement custom algorithms from classical vision to deep learning. Because the training and inference processes happen entirely
. It is specifically designed as a ready-to-use, embedded deep-learning system for end-users who may have little to no experience in programming or image processing. IMAGO Technologies Key Specifications & Features
With dimensions starting at and interfaces like GigE, digital I/Os, and RS-232, the EB is easy to integrate into space-constrained systems. Its drastically reduced data load means a less powerful computing unit can be used, lowering overall system costs.
To help you quickly identify which camera might be right for you, here is a summary of the key specifications: