摘要
国内矿用钢丝绳检测通常采用MT716-2005标准,通过人工目视、触摸等方法来检测钢丝绳表面的损坏情况。人工检测存在因检测人员专业水平参差不齐、人力资源不足或者疲惫而产生错误结果等误判情况。针对该问题,设计了一种基于盘古大模型的矿用钢丝绳表面损毁检测模型。该模型使用NL-means算法去噪和华为云Modelarts平台提供的pangucv-det-protect算法进行物体检测,并对检测结果做出预测,预测数值代表可信度。结果表明,该模型能够有效检测钢丝绳的变形、磨损、锈蚀等损伤情况。
The MT716-2005 standard is usually used for the detection of domestic mining wire rope,and the surface damage of the wire rope is detected by manual visual inspection,touch and other methods.There are misjudgments in manual testing due to uneven professional level of testing personnel,insufficient human resources or fatigue.To solve this problem,a surface damage detection model of mining wire rope based on Pangu large model is designed in this paper.The model uses NL-means algorithm to remove noise and pangu-cv-det-protect algorithm provided by Huawei Cloud Modelarts to detect objects,and makes predictions on the detection results.The predicted values represent credibility.
出处
《工业控制计算机》
2024年第1期1-3,6,共4页
Industrial Control Computer
基金
中央引导地方科技发展资金项目(基础研究项目)(216Z5401G)
河北省高等教育教学改革研究与实践项目(2022GJJG473)
中央高校科研业务费(3142021002)。
