SSL对象检测数据集

SSL对象检测数据集是由伯南布哥联邦大学计算机中心的Roberto Fernandes等人创建的，旨在为RoboCup SSL提供一个开放源代码的基准数据集。该数据集包含931张图像，涵盖了机器人、球和目标三种对象类别，图像来源于不同的光照和角度条件。创建过程中，使用了LabelImg工具进行标注，并将数据随机分为训练集和测试集。该数据集主要用于评估和比较卷积神经网络模型在实时嵌入式系统中的性能，特别是在RoboCup SSL比赛中的应用，以解决快速决策和对象检测的需求。

SSL object detection dataset was created by Roberto Fernandes et al. from the Computer Center of the Federal University of Pernambuco, aiming to serve as an open-source benchmark dataset for RoboCup SSL. The dataset includes 931 images covering three object categories: robot, ball, and goal, which were collected under varying lighting conditions and viewing angles. During the dataset creation process, the LabelImg tool was used for annotation, and the dataset was randomly split into training and test sets. This dataset is primarily used to evaluate and compare the performance of convolutional neural network models in real-time embedded systems, especially for applications in RoboCup SSL competitions, to address the demands of rapid decision-making and object detection.