jan-martens0124/quail_egg

# Quail Egg Detection Dataset

## Overview
This dataset was meticulously curated as part of a school project aimed at developing a machine learning model capable of accurately detecting quail eggs. The motivation stemmed from the practical need to automate the process of egg detection within a henhouse environment. Leveraging both images captured in the henhouse and those obtained from the classroom setting, this dataset represents a diverse range of real-world scenarios.

## Data Collection
The dataset comprises images sourced from multiple environments, including both controlled settings such as classrooms and the dynamic environment of a working henhouse. These images were carefully selected to encompass various lighting conditions, backgrounds, and perspectives to ensure robustness in model training.

## Annotation
All images in this dataset have been meticulously annotated utilizing the free version of the online annotation tool 'Roboflow'. Each quail egg instance within the images is precisely outlined using bounding boxes. Notably, a subset of images features irregular shapes, necessitating polygonal annotations. While traditional training processes often disregard such images, this dataset includes them to enhance model adaptability to real-world complexities.

## Dataset Format
The dataset includes 5776 images.
Eggs-quails are annotated in YOLOv8 format.

The following pre-processing was applied to each image:

The following augmentation was applied to create 3 versions of each source image:
* Equal probability of one of the following 90-degree rotations: none, clockwise, counter-clockwise
* Random shear of between -15° to +15° horizontally and -10° to +10° vertically
* Salt and pepper noise was applied to 1.95 percent of pixels

The following transformations were applied to the bounding boxes of each image:
* Random brigthness adjustment of between -27 and +27 percent

## Usage
Researchers and developers interested in quail egg detection, object detection in dynamic environments, or training YOLOv8 models can leverage this dataset for experimentation and model development. The diversity of environments and annotation complexities makes this dataset invaluable for advancing object detection capabilities in challenging real-world scenarios.

原始信息汇总

Quail Egg Detection Dataset 概述

数据集内容

• 图像数量: 5776张
• 环境多样性: 包含教室和鸡舍等不同环境的图像
• 光照和背景: 涵盖多种光照条件和背景
• 视角多样性: 包含多角度拍摄的图像
• 图像标注: 使用Roboflow工具进行精确的边界框标注，部分图像使用多边形标注以适应不规则形状
• 数据格式: 采用YOLOv8格式进行标注

数据预处理和增强

• 图像旋转: 随机选择90度旋转（无旋转、顺时针、逆时针）
• 图像剪切: 水平方向-15°至+15°，垂直方向-10°至+10°的随机剪切
• 噪声添加: 对1.95%的像素应用盐和胡椒噪声
• 亮度调整: 随机调整-27%至+27%的亮度

应用场景

• 研究与开发: 适用于对鹌鹑蛋检测、动态环境中的物体检测或YOLOv8模型训练感兴趣的研究人员和开发者
• 模型适应性: 数据集的多样性和标注复杂性有助于提升在复杂真实世界场景中的物体检测能力