用轻量神经网络实现实时套管接箍识别，解决井下设备定位难题。
A Neural Network-Based Real-time Casing Collar Recognition System for Downhole Instruments
- 设计专用轻量1D卷积网络，通过深度可分离卷积提升效率
- 模型仅1985参数，F1得分0.972，每秒处理1000次推理
- 适合嵌入式设备部署，满足井下严苛的实时性与功耗要求
套管接箍定位（CCL）测量广泛用于套管井作业中对井下仪器进行精确定位，为射孔等操作提供可靠的深度控制。然而，在井下环境中实现自主接箍识别仍具挑战，因CCL信号常受工具串或套管引起的磁干扰影响，且严格的尺寸与功耗限制制约了复杂算法的应用，特定作业还需实时、原位处理。为此，我们提出针对该领域的轻量级一维卷积神经网络家族——接箍识别网络（CRNs），用于从流式CCL波形中识别接箍特征。通过深度可分离卷积和输入池化优化计算效率，不牺牲精度。最紧凑模型在真实数据上达到0.972的F1分数，仅需1,985个参数和8,208个乘加操作（MACs）。该模型部署于基于ARM Cortex-M7的嵌入式系统，使用TensorFlow Lite for Microcontrollers（TFLM）库，实现每秒1,000次推理，延迟343.2微秒，验证了在严苛井下条件下实现鲁棒、自主、实时接箍识别的可行性。
原文摘要 · Abstract (English)
Casing collar locator (CCL) measurements are widely used as reliable depth markers for positioning downhole instruments in cased-hole operations, enabling accurate depth control for operations such as perforation. However, autonomous collar recognition in downhole environments remains challenging because CCL signals are often corrupted by toolstring- or casing-induced magnetic interference, while stringent size and power budgets limit the use of computationally intensive algorithms and specific operations require real-time, in-situ processing. To address these constraints, we propose Collar Recognition Nets (CRNs), a family of domain-specific lightweight 1-D convolutional neural networks for collar signature recognition from streaming CCL waveforms. With depthwise separable convolutions and input pooling, CRNs optimize efficiency without sacrificing accuracy. Our most compact model achieves an F1-score of 0.972 on field data with only 1,985~parameters and 8,208~MACs, and deployed on an ARM Cortex-M7 based embedded system using TensorFlow Lite for Microcontrollers (TFLM) library, the model demonstrates a throughput of 1,000 inference per second and 343.2 μs latency, confirming the feasibility of robust, autonomous, and real-time collar recognition under stringent downhole constraints.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。