Our new lab is up and running, focusing on the capabilities of the RZ/V2M evaluation board equipped with the Sony IMX568 global shutter image sensor. This platform offers a compelling environment for system designers and developers to explore the world of Vision AI, leveraging the strengths of low-power MCUs, computer vision algorithms and image signal processor.
Streamlined Development Workflow
The RZ/V2M board simplifies software development for various tasks Vision AI processing
Peripherals used in the Lab Setup
- DC Geared motor (12V DC) to rotate the images.
- Chain link and wheel to build conveyor to carry the images.
- Motor driver L298N (12V 2A).
- Power Supply, Oscilloscope, sensors, control ICs, types of adapters/connectors, relays(rating).
- Camera Sensor Input Processing: Pre-process images efficiently captured by the high-resolution Sony IMX568 sensor.
- Low-Power AI Inference: Run AI models on the RZ/V2M’s low-power MCU, optimizing battery life in resource-constrained devices.
- Real-Time Video Streaming: Develop applications that benefit from smooth video streaming capabilities.
Sony IMX568: Global Shutter for Motion Clarity
The system integrates the Sony IMX568, a global shutter image sensor. This sensor technology excels at capturing sharp, distortion-free images, even under rapid motion scenarios.
DRP-AI and YOLOv2: Optimized for IMX568
Our lab utilizes a custom DRP-AI implementation of the compact YOLOv2 application. This application integrates seamlessly with the provided ISP support package, ensuring optimal performance when paired with the IMX568 sensor.
LiveBench: Evaluating RZ/V2M’s Potential
The latest lab on LiveBench features Vision AI on RZ/V2M with IMX568 lab allows for in-depth evaluation capabilities. LiveBench utilizes an external image rotation system, enabling system designers to validate the effectiveness of the Sony IMX568’s global shutter technology with the support of the RZ/V2M platform.
LiveBench offers the following functionalities
- Adjustable Rotation Speed: The image rotation system allows for variations in speed (1-100 rpm), facilitating testing under diverse motion scenarios.
- Live Video Streaming with Rotation: Experience smooth live video streaming with image rotation, powered by the Sony IMX568 sensor.
- Object Detection in Motion: The RZ/V2M platform leverages the Darknet tinyYOLOv2 algorithm for real-time object detection, even during image rotation.
- Pause and Resume Detection: The ability to temporarily pause and resume object detection allows for isolating the impact of global shutter technology.
- Uninterrupted Detection at High Speeds: Observe the continued effectiveness of object detection using the Sony IMX568 sensor, even at high rotational speeds, when the algorithm is paused.
Key Technical Features
- Rapid Object Detection: The Darknet tinyYOLOv2 algorithm ensures fast and accurate object identification.
- Robustness Under Motion Blur: Object detection maintains exceptional precision even with high-speed image rotation.
- Variable Rotation Speeds: Fine-tune the image rotation speed from 1 to 100 rpm to match your specific application requirements.
The launch of this new lab signifies our dedication to advancing research and development in the field of Vision AI.
Vijay Anand, CTO and Co-Founder, TenXer Labs.
We invite system designers, researchers and developers to explore the immense potential of the RZ/V2M evaluation board with the Sony IMX568 sensor and contribute to the creation of applications that leverage the power of computer vision, low-power MCUs and image signal processor.
