DeepSort-3C: Background-Removed Object Dataset for Intelligent Conveyor Belt Classification

An AI-powered object detection and sorting dataset designed for classifying everyday household objects into three categories: Black, Transparent, and Colourful. The dataset includes images of common home items such as toys, containers, bottles, tools, and other domestic objects collected in real-world environments for deep learning–based classification and intelligent conveyor belt routing applications. All images were preprocessed by removing the background of each object to improve feature extraction and increase classification performance. This background elimination helps deep learning models focus more effectively on object properties such as color, texture, and shape while reducing unnecessary environmental noise. The dataset is organized into the following three classes: Colourful objects: 68 images Transparent objects: 24 images Black objects: 46 images This dataset can support research and development in: Object detection Image classification Smart conveyor belt systems AI-based industrial automation Computer vision applications The dataset is particularly useful for training deep learning models intended for automated object sorting and routing systems in industrial and smart manufacturing environments.本数据集为一款面向人工智能的目标检测（object detection）与分拣数据集，旨在将日常家居物品划分为三大类别：黑色（Black）、透明（Transparent）与彩色（Colourful）。数据集收录了真实环境下采集的常见家居物品图像，涵盖玩具、收纳容器、瓶罐、工具及其他家用物件，可应用于基于深度学习（deep learning）的分类任务与智能传送带分拣路径规划场景。 所有图像均经过预处理操作：移除单幅图像中物体的背景，以提升特征提取效率与分类性能。该背景移除步骤可帮助深度学习模型更精准地聚焦于物体的颜色、纹理与形状等核心属性，同时剔除不必要的环境噪声干扰。 本数据集分为以下三个类别： 彩色（Colourful）物品：68张图像 透明（Transparent）物品：24张图像 黑色（Black）物品：46张图像 本数据集可支持以下方向的研发工作： 目标检测（object detection） 图像分类（image classification） 智能传送带系统 基于人工智能的工业自动化 计算机视觉（computer vision）应用 本数据集尤其适用于训练面向工业与智能制造场景下的自动化物体分拣与路径规划系统的深度学习模型。