NIR+RGB datasets
收藏资源简介:
该数据集由澳大利亚数据61研究所的研究人员创建,旨在用于合成近红外(NIR)图像生成和提高水果检测系统的性能。数据集包含从原始nirscene数据集以10、100、200和400的比率过采样得到的127k对图像,以及从SEN12MS卫星多光谱数据集中选出的夏季和全年子集。此外,还包括从商业农场收集的1615对NIR+RGB甜椒图像。数据集的应用领域包括农业领域的植物分割和水果检测,旨在通过提供高质量的数据集来提高深度神经网络模型的泛化能力和部署效率。
This dataset was created by researchers at Data61, Australia, for the purposes of synthetic near-infrared (NIR) image generation and enhancing the performance of fruit detection systems. It contains 127k pairs of images obtained by oversampling the original NIRScene dataset at ratios of 10, 100, 200 and 400, as well as summer and full-year subsets selected from the SEN12MS satellite multispectral dataset. Additionally, it includes 1615 pairs of NIR+RGB sweet pepper images collected from commercial farms. The dataset is applicable to plant segmentation and fruit detection in the agricultural field, aiming to improve the generalization ability and deployment efficiency of deep neural network models by providing high-quality datasets.




