云梯系统儿童单独乘梯分析数据

依据告警数据，及时在轿厢内提醒乘客安全乘梯，设备抓图便于在发生突发状况时，快速了解现场状况，定位电梯及乘客所在位置，对独自乘梯儿童的安全起到实际保障作用；定期对告警数据设备抓图所反映的现场情况进行分析，方便物业对儿童单独乘梯行为进行日常监管，判断是否需要安排现场引导；根据一段时间内告警的数据规模、现场图片、时间等，帮助电梯生产销售企业了解儿童单独乘梯现象情况，选择销售合适的电梯类型，研判是否需要开发定制化的电梯服务，如语音提示，为儿童提供更加便捷、安全的乘梯环境；公开报道一段时间内儿童单独乘梯告警数据，可以提高公众对儿童单独乘梯安全的重视程度。1.通过基于传感器的实时电梯门开关状态检测算法检出处于运行状态的电梯；2.开启电梯内摄像头抓图服务(抓图频率1FPS),并将图片上传到华为云obs。云端算法服务器从华为云obs拉取到图片后，通过基于深度学习神经网络训练的目标检测算法进行图片检测，输出人体相关属性；3.输出的人体属性包含：目标类型category、目标坐标框box、目标结果置信度score、目标匹配groupid、目标年龄识别ageType； 4.对输出的年龄识别进行区分：1代表0-6岁，2代表6岁-12岁，3代表12岁-18岁，4代表18岁-28岁，5代表28岁-40岁，6代表40岁-55岁，7代表55岁以上，当检出标类型中人体仅有1个，且年龄识别为1（0-6岁的人群）时，系统会向平台发送儿童单独乘梯告警，告警时间为时间戳，更加精准。

Based on alarm data, timely safety reminders are provided for passengers taking elevators inside the elevator cab. The device-captured images enable quick understanding of on-site situations and accurate positioning of elevators and passengers during emergencies, effectively ensuring the safety of children riding elevators alone. Regular analysis of on-site conditions reflected by alarm data and captured images allows property management to conduct daily supervision of children's solo elevator riding behavior and determine whether on-site guidance is needed. Additionally, over a given period, by analyzing metrics such as alarm data scale, on-site images, and timestamps, elevator manufacturing and sales enterprises can gain insights into the prevalence of children's solo elevator riding, select appropriate elevator models for sales, and assess whether to develop customized elevator services such as voice prompts to provide a more convenient and safe riding environment for children. Publicizing the alarm data related to children's solo elevator riding over a certain period can also raise public awareness of the safety issues associated with children riding elevators alone.
1. Detect operating elevators using a sensor-based real-time elevator cab door opening and closing state detection algorithm.
2. Activate the in-cab camera image capture service (with a capture frequency of 1 FPS) and upload the captured images to Huawei Cloud OBS. After the cloud-based algorithm server retrieves the images from Huawei Cloud OBS, it conducts image detection via a target detection algorithm trained on deep learning neural networks, and outputs human-related attributes.
3. The output human-related attributes include: target category, target bounding box, target confidence score, target matching group ID, and target age recognition type.
4. Classify the age recognition results as follows: 1 corresponds to children aged 0–6 years, 2 for 6–12 years, 3 for 12–18 years, 4 for 18–28 years, 5 for 28–40 years, 6 for 40–55 years, and 7 for those aged over 55 years. When only one human target is detected and its age recognition type is 1 (0–6 years old), the system will send a children solo elevator riding alarm to the platform, with the alarm time recorded as a timestamp to achieve higher detection accuracy.

搜集汇总

数据集介绍

特点

云梯系统儿童单独乘梯分析数据是一个由浙江新再灵科技股份有限公司自行产生的数据集，包含751条记录，每日更新。该数据集通过深度学习算法检测电梯内儿童单独乘梯行为，应用于安全提醒、物业监管、企业产品开发和公众教育等多个场景。