SWIR Camera Inspection, with Vision AI Trained on Your Images

SWIR Camera Inspection AI Vision Solution

Short-wave infrared (SWIR) covers roughly 900 to 1700 nm in industrial machine vision. Water absorbs SWIR strongly (opens in a new tab), so wet spots and bruises show up dark, and some materials that look opaque in visible light, such as certain packaging or silicon, become partly transparent. That is why SWIR cameras are used for fruit sorting, moisture checks, fill levels through opaque containers, plastic sorting, and silicon wafer inspection.

• LUCID: Triton and Atlas SWIR cameras (opens in a new tab) with Sony SenSWIR sensors (400 to 1700 nm, 0.3 to 5.2 MP), GigE Vision with PoE, and IP67 housings (LUCID cameras with Roboflow)
• JAI: Wave series SWIR area scan and line scan cameras (JAI cameras with Roboflow)

How Roboflow Works with SWIR Cameras

• Collect: SWIR cameras output monochrome images. Upload frames from your line to Roboflow
• Label and train: Label defects on the SWIR images and train a model such as RF-DETR, so it learns what a bruise or wet spot looks like in your infrared images, not in color photos
• Deploy: Run the model with Roboflow Inference on the camera's GigE Vision stream on an edge device such as an NVIDIA Jetson, and send results to your PLC or sorter

‍

Want AI without the integration work? Roboflow's AI1 deployment device (opens in a new tab) packages onboard AI compute, integrated lighting, and an industrial camera into a plug-and-play system for real-time computer vision, with no separate camera, lighting, or edge computer to spec.