README.md

multiple-camera-loader-for-YOLOv5

Up to 4 TIS (The Imaging Source) cameras can be used. Images will be concatenated and fed to model.

• class LoadT4TISCams for tiled TIS camera images
• class LoadV4TISCams for vertically concatenated TIS camera images

Import

  - from cam_loader import LoadT4TISCams, LoadV4TISCams

Also 4 Webcams can be used.

• class LoadT4Streams for tiled Webcam images
• class LoadV4Streams for vertically concatenated Webcam images

Import

- from cam_loader import LoadT4Streams, LoadV4Streams

Add following statement somewhere before detection loop:

dataset = LoadV4TISCams(source, img_size=640, stride=32, auto=True)

You can get image with following statement of detection loop:

for source, frame_lb, frame, rbt_flag, bad in dataset:

Tiled

Vertical

** I've tried FLIR cameras with this logic, and found it doesn't work. Other than TIS and FLIR have not been tested.