Advanced Gas Detection and Classification using MQ Series Sensors Integrated with Machine Learning and Deep Learning Techniques
收藏资源简介：
The dataset focuses on gas detection using MQ series gas sensors in conjunction with machine learning and deep learning models. Data were collected using a setup of MQ gas sensors (MQ-135, MQ-5, MQ-6) connected to an Arduino UNO microcontroller. The sensors were simulated using Proteus software, allowing them to detect gases such as ammonia, carbon dioxide, benzene, natural gas, carbon monoxide, and liquefied petroleum gas (LPG). The data from these sensors were transmitted to a LabView GUI for visualization and storage, and the final dataset was saved in CSV format, containing time-series data with sensor readings and corresponding timestamps. The data collection spanned a 10-day period in March 2023, generating time-series data from six MQ series gas sensors, labeled as Gas1 to Gas6. Each row in the dataset represents a specific timestamp, along with the raw outputs of each sensor and their corresponding gas concentration values in parts per million (PPM). In total, the dataset includes 100,422 samples, each labeled with a "Class" indicator to show the presence or absence of specific gases. The data underwent preprocessing steps, including outlier removal and scaling, to ensure its accuracy and reliability for further analysis.
本数据集聚焦于结合机器学习与深度学习模型，开展基于MQ系列气体传感器（MQ series gas sensors）的气体检测研究。数据采集采用搭载MQ-135、MQ-5、MQ-6型MQ气体传感器，并连接至Arduino UNO微控制器（Arduino UNO microcontroller）的实验装置完成。研究借助Proteus软件对传感器进行仿真，使其可检测氨气、二氧化碳、苯、天然气、一氧化碳及液化石油气（LPG）等多种目标气体。传感器采集的原始数据将传输至LabVIEW图形用户界面（LabVIEW GUI）进行可视化展示与存储，最终数据集以CSV格式保存，包含带传感器读数与对应时间戳的时序数据。 数据采集工作于2023年3月开展，历时10天，共生成6个分别标记为Gas1至Gas6的MQ系列气体传感器的时序数据。数据集中每一行对应一个特定时间戳，同时包含各传感器的原始输出值，以及以百万分比（PPM）为单位的对应气体浓度值。本数据集总计包含100422条样本，每条样本均带有"Class"标签，用于指示特定气体的存在与否。为保障后续分析的准确性与可靠性，数据集已完成异常值剔除、数据缩放等预处理步骤。
