RoseVisuals - Multi-class Dutch Rose Petal Dataset From Controlled Environment.

The RoseVisuals - Multi-class Dutch rose petal dataset represents an extensive advancement in flower image repositories by offering a meticulously curated collection of 1,995 high-resolution, labeled images that capture the significant diversity of Dutch rose petals, encompassing a wide spectrum of color varieties and health conditions. Unlike other datasets that tend to primarily focus on broad plant disease detection or generic flower classification, often neglecting the importantance of petal-level features -RoseVisuals addresses this critical gap and offers a broader view, capturing a wide range of petal colors and their health conditions. The dataset comprises images of petals in shades like red, yellow, pink, purple, white, orange, plus stunning bi-color and multi-color varieties. On top of that, it sorts them by health—whether they’re fresh, dry, or diseased.
In total, RoseVisuals zip folder includes 1,995 images in jpg format. Breaking it down by color, there are 244 Bi-color, 116 Multi-color, 211 Orange, 239 Pink, 237 Purple, 230 Red, 214 White, 239 Yellow. As for health, 828 are Fresh petals, 902 are Dry petals, and 265 are Diseased petals. All images in the RoseVisuals dataset were captured under controlled lighting conditions and using the Samsung S21 Ultra’s pro camera mode, ensuring uniformity and clarity. The built-in device flash was used for all images to minimize shadows and enhance detail, and every image was taken at a consistent resolution of 3000×4000 pixels with portrait frame. Each image has been pre-processed with a unique ID—like "FRESH_RED_001," where "FRESH" shows the petal’s condition, "RED" is the color, and "001" is a three-digit identifier to keep everything organized. This rich annotation assists deep learning researchers to explore fine-grained visual cues related to petal health, color variation, and disease manifestation. This dataset is perfect for training machine learning models, especially for quality checks or projects tied to agriculture, perfumes, edible products, flavors, or even cosmetics. It’s a treasure trove for anyone looking to explore the world of Dutch roses!

RoseVisuals——多分类荷兰玫瑰花瓣数据集是花卉图像库领域的重要进展，其精心收录了1995张高分辨率标注图像，全面展现了荷兰玫瑰花瓣的丰富多样性，涵盖广泛的色彩品类与健康状态。相较于其他大多聚焦于广谱植物病害检测或通用花卉分类、却往往忽视花瓣级特征重要性的数据集，RoseVisuals填补了这一关键研究空白，其采集的图像覆盖了更全面的花瓣色彩与健康状态范畴。
数据集包含红色、黄色、粉色、紫色、白色、橙色等单色花瓣图像，以及极具特色的双色与多色品种。在此基础上，数据集还按照健康状态对图像进行分类：新鲜花瓣、干枯花瓣与染病花瓣。
RoseVisuals的压缩包总计包含1995张JPG格式图像。按色彩分类统计：双色花瓣图像244张、多色花瓣图像116张、橙色花瓣211张、粉色花瓣239张、紫色花瓣237张、红色花瓣230张、白色花瓣214张、黄色花瓣239张。按健康状态分类：新鲜花瓣828张、干枯花瓣902张、染病花瓣265张。
所有数据集图像均在可控光照条件下采集，使用Samsung S21 Ultra专业相机模式拍摄，以确保图像的一致性与清晰度。拍摄过程中启用了设备内置闪光灯，以减少阴影、增强细节，所有图像均采用统一的3000×4000像素分辨率与竖幅构图。
每张图像均已完成预处理并分配唯一标识符，格式为"FRESH_RED_001"，其中"FRESH"代表花瓣健康状态，"RED"代表花瓣色彩，"001"为三位数字编号以实现系统化管理。这种丰富的标注信息可帮助深度学习研究人员探索与花瓣健康状态、色彩变化及病害表现相关的细粒度视觉线索。
该数据集非常适合用于训练机器学习模型，尤其适用于农业、香水、食用产品、香精香料乃至化妆品相关的质量检测类项目。对于任何致力于探索荷兰玫瑰领域的研究者而言，本数据集都是一座宝贵的资源宝库。