油气钻采井场情况复杂且安全隐患种类多，为提高井场安全隐患识别的精确性，提出一种基于改进YOLOv5的油气钻采井场安全隐患识别方法。首先，为解决图片背景复杂，识别难度加大的问题，在骨干网络引进SimAM注意力机制；其次，为解决隐患类型的尺度差异较大且一张图片中存在多种尺度的问题，用自适应空间特征融合（ASFF）替换原有特征融合方式；最后，通过与其他模型进行比对来验证改进模型的隐患识别效果。结果表明：改进后的YOLOv5模型识别平均精确值提高了10.4%，对油气钻采井场安全隐患具有较好的识别效果。为解决油气钻采井场安全隐患视频监控识别的限制，研发一套智能穿戴设备，有效提高了井场安全隐患识别的便携性。

Oil and gas drilling and production wellsites are complex and have many types of potential safety hazards, in order to improve the accuracy of the identification of potential safety hazards in wellsites, an oil and gas drilling and production wellsite potential safety hazard identification method based on improved YOLOv5 was proposed. Firstly, in order to solve the problem that the background of the picture was complex and the recognition difficulty increased, the SimAM attention mechanism was introduced in the backbone network; secondly, in order to solve the problem that the scales of the types of hidden hazards were different and there were multiple scales in one picture, the original feature fusion was replaced by adaptive spatial fusion of features (ASFF). Lastly, the hidden hazard recognition effect of the improved model was validated by comparing the model with other models. The results show that the improved YOLOv5 model improves the average accuracy value of recognition by 10.4%, and has a better recognition effect on the safety hazards of oil and gas drilling and production well sites. In order to solve the limitation of video monitoring and identification of oil and gas drilling and production wellsite safety hazards, a set of intelligent wearable device was developed, which effectively improved the portability of the identification of wellsite safety hazards.