Research and Optimization of Tubular Column Joint Identification Technology for Intelligent Drilling Platforms
Open Access
- 29 February 2024
- journal article
- Published by Madison Academic Press in Advances in Engineering Technology Research
- Vol. 10 (1) , 191
Abstract
An improved target detection algorithm based on YOLOv5s is proposed to be applied to identify and localize tubular column joints in The improved SC-YOLOv5s model is obtained by combining SPPF, CBAM and the improved multi-scale feature fusion network BiFPN. The improved SC-YOLOv5s model is obtained by combining SPPF, CBAM and the improved multi-scale feature fusion network BiFPN. The SPPF structure introduces spatial pyramid pooling in the network, which can effectively capture the features of the target at different scales and improve detection precision. different scales and improve the detection precision and recall of the target, while the CBAM attention mechanism can adaptively learn the target's spatial and channel feature relationships. s spatial and channel feature relationships, improving the model's ability to distinguish between targets and anti-interference ability The improved multi-scale feature fusion network BiFPN combines Bottom-up and Top-down feature transfer mechanisms, which can better fuse feature information at different levels, making the model's ability to distinguish between targets and anti-interference ability. information at different levels, making the model better adapted to detection tasks in various complex environments. The experimental results show that the improved algorithm achieves the following results that the improved algorithm achieves significant performance improvement in the pipe-column joint detection task, with a mAP value of 99.44%, a frame rate of 136 FPS, and an accuracy of 98.4%, which is 6.6% higher than the original model. The algorithm also improves sensing capability, accuracy and robustness, and outperforms the original model. The algorithm also improves sensing capability, accuracy and robustness, and outperforms other mainstream models in terms of precision. This provides a reference for automatic processing of tubular columns in intelligent oil rigs.Keywords
This publication has 8 references indexed in Scilit:
- YOLOv5s-CBAM-DMLHead: A lightweight identification algorithm for weedy rice (Oryza sativa f. spontanea) based on improved YOLOv5Crop Protection, 2023
- SAR ship detection based on improved YOLOv5 and BiFPNICT Express, 2023
- Tomato recognition and location algorithm based on improved YOLOv5Computers and Electronics in Agriculture, 2023
- Intelligent identification of metal corrosion based on Corrosion-YOLOv5sDisplays, 2023
- Machine learning for drilling applications: A reviewJournal of Natural Gas Science and Engineering, 2022
- An improved YOLOv5-based vegetable disease detection methodComputers and Electronics in Agriculture, 2022
- Underwater scallop recognition algorithm using improved YOLOv5Aquacultural Engineering, 2022
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016