OD_AVI: An object detection in the around view image dataset for the AVP

The repository containing the OD_AVI dataset, which can be downloaded at OD_AVI. The OD_AVI comprises 9828 fisheye images collected from the indoor and outdoor parking lots. For each image, the 2D bounding box and ground lines of objects are carefully labeled by the Labelme.

Dataset Contents

This dataset version consists of 9828 fisheye images with annotations and the intrinsic and extrinsic parameters for the AVM.

• RGB images
• 2D bounding boxes
• Ground Lines
• Intrinsic parameters for fisheye camea
• Extrinsic parameters for AVM

Citation

@article{Wei Li,
	title = {Object ground lines regression and mapping from fisheye images to around view image for the AVP},
	date = {2021-07-14}
}