Upload folder using huggingface_hub

uavdet-small/README.dataset.txt

@@ -0,0 +1,80 @@
+# uavdet-small-txtvh-ysli > 2025-03-13 7:42am
+https://universe.roboflow.com/rf-100-vl/uavdet-small-txtvh-ysli
+
+Provided by a Roboflow user
+License: MIT
+
+# Overview
+
+- [Introduction](#introduction)
+- [Object Classes](#object-classes)
+  - [Bicycle](#bicycle)
+  - [Bus](#bus)
+  - [Car](#car)
+  - [Human](#human)
+  - [Motorbike](#motorbike)
+  - [Truck](#truck)
+  - [Van](#van)
+
+# Introduction
+
+This dataset aims to provide annotations for various objects in urban environments as captured from an aerial perspective. The task is to solve object detection for city planning and traffic analysis. The dataset consists of the following classes:
+
+- **Bicycle:** Pedal-powered vehicles used for transportation.
+- **Bus:** Large public transport vehicles designed to carry many passengers.
+- **Car:** Common personal vehicles used for private transportation.
+- **Human:** Individuals walking or standing.
+- **Motorbike:** Two-wheeled motor vehicles.
+- **Truck:** Larger vehicles for transporting goods.
+- **Van:** Medium-sized vehicles used for goods or passenger transport.
+
+# Object Classes
+
+## Bicycle
+### Description
+Bicycles are typically small, two-wheeled vehicles with a visible frame, wheels, and handlebars.
+
+### Instructions
+Draw bounding boxes around visible bicycles. Ensure the box includes the entire bicycle frame and the wheels. Do not annotate if less than 20% of the bicycle is visible. Ignore extremely blurry objects or those below the size threshold.
+
+## Bus
+### Description
+Buses are elongated vehicles with multiple windows and are larger than other vehicles on the road.
+
+### Instructions
+Enclose the entire bus within the bounding box, including all visible windows and body sections. Do not include other vehicles or objects touching the bus. Annotate buses even if partially obscured as long as they are clearly recognizable.
+
+## Car
+### Description
+Cars are smaller than buses and trucks, with distinct shapes often including four wheels and a streamlined body.
+
+### Instructions
+Annotate all visible parts of the car, ensuring the box encompasses the body and wheels. Avoid labeling if the car is unidentifiable due to significant obstructions. Ensure not to confuse vans with cars based on their size and shape.
+
+## Human
+### Description
+Humans are the smallest objects in this dataset, typically on foot, visible on crosswalks, sidewalks, or other pedestrian zones.
+
+### Instructions
+Draw a tight bounding box around individual humans, including their entire body when visible. Do not label if the individual is obscured to an unrecognizable extent or is below the size threshold in the image.
+
+## Motorbike
+### Description
+Motorbikes are two-wheeled vehicles, similar in size to bicycles but with a more robust frame and often seen with passengers.
+
+### Instructions
+Capture the full extent of the motorbike, ensuring visibility of the wheels and general frame. Do not include irrelevant parts or nearby objects unless critical to the recognition of the motorbike.
+
+## Truck
+### Description
+Trucks are larger vehicles designed for transporting goods, easily distinguishable by their size and often visible in construction areas.
+
+### Instructions
+Annotate the entire truck, capturing both the cab and trailer sections if visible. Ensure the bounding box is not excessively large to include unrelated objects. Disambiguate from buses and vans by checking for cargo areas.
+
+## Van
+### Description
+Vans are medium-sized vehicles, typically larger than cars but smaller than buses and trucks, often used for transport of goods or people.
+
+### Instructions
+Include the whole van within the annotation, observing the body structure and size. Differentiate from cars by focusing on the van's heightened body and rear space. Do not include adjacent objects unless they obscure part of the van critical for identification.

uavdet-small/README.roboflow.txt

@@ -0,0 +1,27 @@
+
+uavdet-small-txtvh-ysli - v1 2025-03-13 7:42am
+==============================
+
+This dataset was exported via roboflow.com on March 14, 2025 at 6:09 AM GMT
+
+Roboflow is an end-to-end computer vision platform that helps you
+* collaborate with your team on computer vision projects
+* collect & organize images
+* understand and search unstructured image data
+* annotate, and create datasets
+* export, train, and deploy computer vision models
+* use active learning to improve your dataset over time
+
+For state of the art Computer Vision training notebooks you can use with this dataset,
+visit https://github.com/roboflow/notebooks
+
+To find over 100k other datasets and pre-trained models, visit https://universe.roboflow.com
+
+The dataset includes 1000 images.
+Uavdet-small-txtvh-ysli are annotated in COCO format.
+
+The following pre-processing was applied to each image:
+
+No image augmentation techniques were applied.
+
+