The invention discloses an artificial intelligence inspection and detection method for an elevator, and relates to a video target tracking technology. The method comprises the following two steps: 1, an offline training stage: 1, preparing data; 2, feature extraction; 3, model training; 2, an online application stage: 1, initializing; 2, target tracking; 3, updating and correcting; 4, outputting a judgment result; according to the method, the action process of the elevator dynamic inspection and detection items is tracked, then recognition and judgment are carried out, and finally wanted judgment information is output. For different action processes, tracked target objects are different, extraction algorithms of the target objects are different, feature extraction algorithms are different, and tracking models are different.
本发明公开了一种电梯人工智能检验检测方法,涉及到了视频目标跟踪技术。该方法分为两大步骤,第一步离线训练阶段:1.数据准备;2.特征提取;3.模型训练;第二步在线应用阶段:1.初始化;2.目标跟踪;3.更新与纠正;4.输出判定结果;本发明通过对涉及电梯动态检验检测项目的动作过程进行跟踪,然后进行识别和判定,最后输出想要的判定信息。对于不同的动作过程,所跟踪的目标物体不同,目标物体的提取算法不同,特征提取算法不同,跟踪模型不同。
Artificial intelligence inspection and detection method for elevator
一种电梯人工智能检验检测方法
2024-08-13
Patent
Electronic Resource
Chinese
Artificial intelligence inspection and detection method for elevator
European Patent Office | 2024
|Artificial intelligence inspection and detection device for elevator shaft data
European Patent Office | 2024
|Artificial intelligence inspection and detection device for elevator machine room data
European Patent Office | 2024
|Artificial intelligence inspection and detection device for comprehensive data of elevator landing
European Patent Office | 2024
|