Local inference
See an edge prototype running
This demo shows apple detection on a portable Raspberry Pi 5 and Hailo prototype. The Abraia Vision SDK runs inference on images and video and composes stages for detection, tracking, line crossing, and time spent inside a region.
The SDK supports ONNX inference and Hailo acceleration for compatible models and hardware. The on-screen performance is specific to this device, model, and recording.
Recorded inference demos
Compare tomato and apple detection runs
Sample on-screen FPS readings in these recordings are about 26 FPS with Hailo HEF and 4 FPS with ONNX for tomato detection, and 14 FPS with Hailo HEF and 3 FPS with ONNX for apple detection. Actual throughput depends on the model, device, input resolution, and runtime.
Apple detection
These are run-specific readings from demonstration videos, not a controlled benchmark. For a deployment comparison, test the same footage, resolution, model settings, and target device.
These clips show standard color-image detection. For multispectral projects, Vision Studio supports spectral analysis and workflow prototyping; validate sensor input and the deployment path for each target setup. See multispectral imaging.
Use cases
Select the task before the hardware
A project begins with the decision the system must make and how results will be checked.
People flow and queues
Detect and track people, count crossings, or measure tracked time within a defined region.
Explore queue monitoring →Inspection and counting
Detect or segment objects in camera feeds, then apply task-specific counting or quality logic.
Explore agriculture →Deployment process
Prove the system with representative data
Start with sample footage from the intended camera and lighting conditions. Choose an input format, model, processing stages, and target device; then test accuracy, latency, and failure cases in the actual environment.
Vision Studio supports pipeline prototyping. Device deployment is an engineering step scoped to the selected workload. The SDK provides the Python building blocks.
Bring your vision problem to us
Tell us about the images, sensor, operating environment, and decision you want the system to support. We can identify a practical first evaluation.