CottonInsect: An image dataset for cotton field insect identification study

To solve the problem of insect classification and detection in cotton fields under complex backgrounds, the cotton field insect recognition research image dataset is constructed. This dataset contains 13 common cotton field insects, with a total of 3225 images and a total of 24GB of original images. These image data were collected from different times, locations, and environments, including images of cotton field insects in different postures and growth stages in the real environment. Datasets are manually filtered and annotated into two formats: image classification and object detection. Ultimately, it can provide basic data for tasks such as insect image classification and object detection in cotton fields. 

为解决复杂背景下棉田昆虫分类与检测的研究需求与实际难题，本研究构建了棉田昆虫识别科研图像数据集。该数据集涵盖13种常见棉田昆虫，总计包含3225张图像，原始图像总存储容量达24GB。这批图像数据采集自不同时间、地点与环境场景，涵盖真实田间环境下不同姿态、不同生长阶段的棉田昆虫图像。本数据集经过人工筛选与标注，分为图像分类（image classification）、目标检测（object detection）两种标注格式。最终可为棉田昆虫图像分类、目标检测等相关任务提供基础数据支撑。

搜集汇总

背景与挑战

背景概述

CottonInsect是一个用于棉田昆虫识别研究的图像数据集，包含13种常见昆虫的3225张图像（24GB），涵盖不同环境和生长阶段。数据集经过人工标注，适用于图像分类和目标检测任务，并获得了国家级科研项目支持。