2022年用于自动收获和可视化产量估计的水果图像数据集

果品日间数据集是第一个可用于果实采摘前和采摘中应用的公开数据集。该数据集在天然果园环境中获取，并分为两个不同子集，可用于不同采摘阶段的应用。第一个子集由8079 张jpg 格式224×224像素彩色图像组成，由于不同的成像角度和尺度、不同的光照、不同的水果品种和成熟阶段，以及不同的水果套袋状态，收集到的图像显示出丰富变化。上述图像根据水果品种、成熟度和采摘情况被标记为不同的类别，已用于评估水果分类的深度学习算法。第二个子集包含采摘期间获得的图像、视频和重量测量值。该数据集可用于果实产量预测相关研究。https://doi.org/10.21227/x46j-sk98

The Daytime Fruit Dataset is the first public dataset applicable to pre-harvest and in-harvest fruit picking scenarios. Collected in natural orchard environments, this dataset is divided into two distinct subsets tailored for different harvesting stage applications. The first subset comprises 8079 JPEG-format 224×224 pixel color images. The collected images exhibit substantial variations due to diverse imaging angles and scales, varying lighting conditions, different fruit varieties, ripening stages, and fruit bagging states. These images are annotated into distinct categories based on fruit variety, ripeness and harvesting conditions, and have been utilized to evaluate deep learning algorithms for fruit classification. The second subset contains images, videos and weight measurements acquired during the harvesting period. This dataset can support research related to fruit yield prediction. https://doi.org/10.21227/x46j-sk98