大型船闸人字闸门门缝错位图像识别方法研究

Research on image recognition method for misalignment of gates in large ship locks

摘要

探讨了基于图像识别技术的某大型船闸人字闸门门缝错位检测方案。采用高清晰度图像采集设备，并通过对采集到的图像进行预处理、特征提取和支持向量机分类器训练等步骤，实现对人字闸门门缝、错位的实时检测。通过现场样机测试和误差分析，证明该方案具有实用性和良好的检测效果，满足该船闸的实际应用要求。

Abstract

The misalignment detection scheme for the miter gate of a large ship lock using image recognition technology was discussed. High-definition image acquisition equipment was used. Real-time detection of gaps and dislocations in the miter gate was achieved through pre-processing, feature extraction, and SVM classifier training. Field prototype testing and error analysis have demonstrated the practicality and effectiveness of the scheme, meeting the practical application requirements of ship locks.

关键词

门缝错位检测 / 灰度化 / 亚像素迭代 / 姿态评估

Key words

gate joint dislocation detection / graying / subpixel iteration / attitude evaluation