生产现场违规事项视觉检测数据

该数据由视觉检测系统抓取生产现场画面所得，主要用来捕捉生产现场上对于人员手持物料、物料掉落或未按规范放置等违规行为，如发现违规行为会马上产生报警，可以及时制止其违规行为，保证车间生产安全和生产效率。还可分析违规行为，对车间人员开展相应的安全教育，以加强车间安全保障。1. 数据采集与处理：
- 使用视觉传感器和图像处理技术实时监测生产区域。
2. 违规检测算法：
- 基于深度学习模型，识别生产区域中的人员和物料。
- 检测人员携带物料、物料掉落和放置不规范等违规行为。
- 利用机器学习算法分析图像，识别违规行为模式，如手持物料数量、盘中物料数量、物料掉落数量、正面开门数量、手持物料的时间和位置、物料掉落的频率等。
3. 报警与处理：
- 根据视觉传感器抓拍现场照片，通过图像处理技术识别照片中手持物料数量、物料掉落数量、正面开门数量、手持物料的时间和位置、物料掉落的频率，其中物料掉落的数量应低于1个，否则报警，正面开门数量应低于2，否则报警，手持物料的时间应小于20s，否则报警，手持物料的位置应为边缘，否则报警，物料掉落的频率应小于1个/8h（物料掉落的频率计算的是一个班时内的频率），否则报警。
实时生成违规报警，并将报警信息传输至监控中心或相关人员，报警信息为输出0则表示不报警，输出1则表示报警， 并且记录报警时间。
- 提供实时数据和分析结果，支持管理人员迅速响应和处理违规事件。

This dataset is collected from production workshop scenes captured by visual inspection systems. It is primarily used to detect violation behaviors in production workshops, such as personnel holding materials, material dropping, or improper material placement. Once a violation is detected, an immediate alarm will be triggered to stop the violation in time, ensuring workshop production safety and production efficiency. Additionally, it can analyze violation behaviors to conduct targeted safety education for workshop personnel, thereby enhancing workshop safety guarantees.
1. Data Collection and Processing:
- The production area is monitored in real time using visual sensors and image processing technologies.
2. Violation Detection Algorithm:
- Deep learning models are employed to identify personnel and materials within the production area.
- Detect violation behaviors including personnel carrying materials, material dropping, and improper material placement.
- Machine learning algorithms are used to analyze images and identify violation behavior patterns, such as the quantity of held materials, quantity of materials in the tray, quantity of dropped materials, quantity of front-facing door openings, holding duration and position of materials, and material dropping frequency, etc.
3. Alarm and Handling:
- On-site photos captured by visual sensors are analyzed via image processing technologies to identify the above-mentioned parameters. Alarms will be triggered if any of the following thresholds are violated: the quantity of dropped materials ≥ 1, the quantity of front-facing door openings ≥ 2, the holding duration of materials ≥ 20 seconds, the holding position of materials is not at the edge, and the material dropping frequency ≥ 1 per 8 hours (the frequency is calculated within one work shift). Violation alarms are generated in real time and transmitted to the monitoring center or relevant personnel. The alarm information uses output 0 to indicate no alarm and output 1 to indicate alarm, and the alarm time is recorded.
- Real-time data and analysis results are provided to support managers in quickly responding to and handling violation incidents.

搜集汇总

数据集介绍

特点

该数据集包含566条实时更新的生产现场违规行为检测数据，通过视觉检测系统和深度学习模型识别人员手持物料、物料掉落等违规行为，并实时生成报警信息，用于保障车间生产安全和效率。