Road Defects 5

## Overview

Road Defects 5 is a dataset for semantic segmentation tasks - it contains Road Erros annotations for 3,073 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

🎯 Pothole Detection for Road Safety: A Segmentation Dataset

🛑 Importance of Pothole Detection:

Pothole detection is crucial as it significantly contributes to road safety, maintenance efficiency, and cost-effectiveness. By identifying and repairing potholes promptly, we can prevent accidents, save on repair costs, and ensure smooth traffic flow.

🔍 Dataset Details:

This dataset was sourced via roboflow.com and contains a total of 780 images annotated in the YOLOv8 format for pothole detection and segmentation. The images have undergone pre-processing and augmentation to ensure they are primed for training robust models.

🔄 Pre-processing applied to each image includes:

• Auto-orientation of pixel data (with EXIF-orientation stripping)
• Resizing to a dimension of 640x640 (Stretch)

🎭 Augmentation was performed to create 3 versions of each source image, exclusively on the training data, comprising:

• 50% probability of horizontal flip
• Random cropping (0 to 20% of the image)
• Random rotation (-15 to +15 degrees)
• Random shearing (-5° to +5° horizontally and vertically)
• Random brightness adjustment (-25 to +25 percent)
• Random exposure adjustment (-25 to +25 percent)

🖼️ Train Set: 720 images 🔍 Validation Set: 60 images

By utilizing this dataset, you are not just developing a pothole detection model, but contributing to a cause that ensures safer and better-maintained roads. Get started on this fulfilling journey towards creating impact-driven computer vision models for real-world challenges!

Road Defects Group 3

## Overview

Road Defects Group 3 is a dataset for instance segmentation tasks - it contains Road Errors annotations for 5,478 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

The current highway waveform guardrail recognition technology has encountered problems with low segmentation accuracy and strong noise interference. Therefore, an improved U-net semantic segmentation model is proposed to improve the efficiency of road maintenance detection. The model training is guided by mixed expansion convolution and mixed loss function, while the presence of guardrail shedding is investigated by using partial mean values of gray values in ROI region based on segmentation results, while the first-order detail coefficients of wavelet transform are applied to detect guardrail defects and deformation. It has been determined that the Miou and Dice of the improved model are improved by 8.63% and 17.67%, respectively, over the traditional model, and that the method of detecting defects in the data is more accurate than 85%. As a result of efficient detection of highway waveform guardrail, the detection process is shortened and the effectiveness of the detection is improved later on during road maintenance.

Road Defects 4

## Overview

Road Defects 4 is a dataset for instance segmentation tasks - it contains Road Errors annotations for 4,672 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Proportion of different scene categories in the data.

Line Defect Detection

## Overview

Line Defect Detection is a dataset for instance segmentation tasks - it contains Lines annotations for 204 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

Semantic Segmentation Roads

## Overview

Semantic Segmentation Roads is a dataset for semantic segmentation tasks - it contains Roads annotations for 341 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).