Bird Observation Dataset, Yogyakarta 2025

The Bird Observation Dataset is a collection of original bird images acquired from a private data collection campaign conducted in Yogyakarta City, Indonesia, during November 2025. The dataset was developed to support research in bird detection, visual recognition, and object differentiation in outdoor environments, particularly for applications related to aerial surveillance and computer vision. All images were captured directly by the authors using DSLR cameras, ensuring high-quality visual data under natural imaging conditions. Image acquisition was carried out during three different time periods—daytime, late afternoon, and nighttime—to capture variations in illumination, background appearance, and visibility. This multitemporal acquisition strategy enables realistic representation of outdoor lighting transitions, including strong daylight, sunset conditions, and low-light nighttime scenes. The dataset includes diverse bird appearances with variations in orientation, scale, flight posture, and distance from the camera, as well as complex backgrounds such as urban structures, vegetation, and open sky typical of the Yogyakarta area. As the images were collected in real-world environments without artificial scene manipulation, the dataset reflects natural environmental complexity, including motion blur and background clutter. This dataset can be used independently or in combination with drone datasets to support research on false-positive reduction in drone detection systems, bird classification, and aerial object discrimination. The dataset is intended for academic and non-commercial research purposes.

The dataset is organized into image folders according to acquisition time periods and annotation availability:

1. Birds_01>>Images_Daytime/ — Bird images captured under daylight conditions
2. Birds_02>>Images_Afternoon/ — Bird images captured during late afternoon or sunset
3. Birds_03>>Images_Night/ — Bird images captured under nighttime illumination
4. Annotations/ — Annotation files (if applicable), provided in standard computer vision formats
5. Documentation/ — Dataset description and usage notes

All image files are provided in standard image formats (e.g., JPG or PNG) suitable for direct use in computer vision pipelines.

鸟类观测数据集（Bird Observation Dataset）为原创鸟类图像合集，采集自2025年11月于印度尼西亚日惹市开展的私人数据采集项目。本数据集旨在支撑户外环境下的鸟类检测、视觉识别与目标区分相关研究，尤其适配航空监测与计算机视觉领域的应用需求。所有图像均由研究团队成员使用数码单反相机（DSLR）直接拍摄，确保自然成像条件下的高质量视觉数据。图像采集覆盖三个不同时段：日间、傍晚及夜间，以捕捉光照、背景形态与可见度的差异。这种多时段采集策略可真实还原户外光照变化场景，包括强光日光、日落环境以及低光照夜间场景。本数据集涵盖多样化的鸟类外观形态，包含拍摄角度、尺度、飞行姿态以及与相机距离的差异，同时搭配日惹地区典型的复杂背景，如城市建筑、植被与开阔天空。由于所有图像均采集自真实户外环境，未经过人工场景干预，本数据集可反映真实的环境复杂性，包括运动模糊与背景杂乱等情况。本数据集可独立使用，或与无人机数据集结合使用，以支撑无人机检测系统误报抑制、鸟类分类以及航空目标区分相关研究。本数据集仅可用于学术与非商业研究用途。

本数据集按照采集时段与标注可用性分为以下图像文件夹：
1. Birds_01>>Images_Daytime/ — 日间光照条件下拍摄的鸟类图像
2. Birds_02>>Images_Afternoon/ — 傍晚或日落时段拍摄的鸟类图像
3. Birds_03>>Images_Night/ — 夜间光照条件下拍摄的鸟类图像
4. Annotations/ — 标注文件（如适用），采用计算机视觉领域标准格式提供
5. Documentation/ — 数据集说明与使用须知

所有图像文件均采用标准图像格式（如JPG或PNG）提供，可直接用于计算机视觉处理流程。