Dataset of Annotated Rice Panicle Image from Bangladesh

This dataset focuses on drone-based rice panicle detection in Gazipur, Bangladesh, offering valuable visual data to researchers in agricultural studies. Captured using an advanced drone with a 4K resolution camera, the dataset comprises 2193 high-resolution images of rice fields and 5701 images after augmentation. All the images are annotated with precision to aid in automated rice panicle identification. Its main purpose is to support the development of algorithms and systems for critical agricultural tasks like crop monitoring and yield estimation, as well as disease identification and plant health evaluation. The dataset's creation involved extracting frames from drone-recorded video footage and meticulously annotating them with manual and deep learning algorithms using a semi-automatic approach.

本数据集聚焦孟加拉国加济布尔地区基于无人机（drone）的稻穗（rice panicle）检测任务，可为农业研究领域的科研人员提供极具价值的视觉数据。本数据集采用配备4K分辨率相机的先进无人机采集，包含2193张稻田高分辨率原始图像，以及经数据增强处理后的5701张图像。所有图像均经过高精度标注，可助力自动化稻穗识别任务。本数据集的核心用途在于支撑面向关键农业任务的算法与系统研发，相关任务涵盖作物监测、产量估算、病害识别以及植株健康评估等。本数据集的构建流程包括从无人机录制的视频素材中提取帧图像，并采用半自动方式结合人工标注与深度学习算法完成精细标注。

搜集汇总

背景与挑战

背景概述

该数据集是一个基于无人机采集的孟加拉国水稻穗图像数据集，包含2193张原始高分辨率图像和5701张增强后图像，所有图像均经过精确标注。其主要目的是支持农业研究中的作物监测、产量估算和疾病识别等算法开发，采用半自动标注方法结合手动和深度学习技术。