FruitSeg30_Segmentation Dataset & Mask Annotations

The “FruitSeg30_Segmentation Dataset & Mask Annotations” is a comprehensive collection of high-resolution images of various fruits, accompanied by precise segmentation masks. We structured this dataset into 30 distinct classes, which containing 1969 images and their corresponding masks, with each measuring 512×512 pixels. Each class folder contains two subfolders: “Images” with high-quality JPG images captured under diverse conditions and “Mask” with PNG files representing the segmentation masks. We meticulously collected the dataset from various locations in Malaysia, Bangladesh, and Australia, ensuring a robust and diverse collection suitable for training and evaluating image segmentation models like U-Net. This resource is ideal for automated fruit recognition and classification applications, agricultural quality control, and computer vision and image processing research. By providing precise annotations and a wide range of fruit types, this dataset serves as a valuable asset for advancing research and development in these fields."FruitSeg30_分割数据集与掩码标注集"（FruitSeg30_Segmentation Dataset & Mask Annotations）是一套囊括多样水果高分辨率图像的综合性数据集，配套高精度分割掩码。我们将该数据集划分为30个独立类别，共计包含1969张图像及其对应掩码，所有图像与掩码的分辨率均为512×512像素。每个类别的文件夹下设两个子文件夹：其一为存储多场景下高质量JPG图像的"Images"子文件夹，其二为存储分割掩码的PNG格式"Mask"子文件夹。本数据集采集自马来西亚、孟加拉国与澳大利亚的多个采集点位，采集过程严谨细致，确保了数据集的多样性与鲁棒性，可用于训练与评估U-Net等图像分割模型。该数据集非常适用于自动化水果识别与分类应用、农业质量管控，以及计算机视觉与图像处理领域的研究工作。凭借精准的标注与丰富的水果品类覆盖，本数据集可作为推动上述领域研究与开发的宝贵资源。

搜集汇总

背景与挑战

背景概述

该数据集是一个包含30类水果的高分辨率图像分割数据集，共1969张512×512像素的图像及对应掩码，结构清晰分为图像和掩码文件夹。数据收集自马来西亚、孟加拉国和澳大利亚，旨在支持图像分割模型训练、水果自动识别和农业质量控制等计算机视觉应用，具有多样性和实用性。