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技术难题，本研究构建了棉田昆虫识别研究图像数据集。该数据集包含13种常见棉田昆虫，共计3225张图像，原始图像总容量达24GB。这些图像数据采集自不同时间、地点与环境，涵盖真实场景下不同姿态、不同生长阶段的棉田昆虫图像。数据集经人工筛选与标注，分为图像分类与目标检测两种格式，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。