预置位自动校准纠偏方法及系统的研究与实现OA
Research and Implementation of an Automatic Calibration and Deviation Correction Method and System for Preset Positions
为解决云台摄像机因长期运行及频繁调用预置位而产生的位置偏移问题,文章提出了一种预置位自动校准纠偏方法.该方法首先获取清晰度满足预设阈值的基准图像与待检测图像,进而分别采用特征点匹配、模板匹配及目标检测算法计算各自的偏移量及其匹配置信度,并基于非线性权重融合策略综合各偏移量,得到最终修正值,实现自动纠偏.系统无须人工干预即可完成全流程校准.实验结果表明,该方法在多种监控场景下能有效降低预置位偏移误差,显著提高图像定位精度与目标识别准确率,且具有良好的实时性和工程应用价值.
To solve the problem of position deviation caused by long-term operation and frequent recall of preset positions in pan-tilt cameras,this paper proposes an automatic calibration and deviation correction method for preset positions.The method first obtains a reference image and an image to be detected that satisfy a preset clarity threshold.It then respectively employs feature point matching,template matching,and object detection algorithms to calculate the corresponding offsets and matching confidence scores.Based on a nonlinear weight fusion strategy,the method fuses the offsets to obtain the final correction value,thereby achieving automatic deviation correction.The system can complete the entire calibration process without manual intervention.The experimental results show that the method can effectively reduce preset-position deviation errors in various surveillance scenarios,significantly improve image positioning accuracy and target recognition accuracy,and demonstrate good real-time performance and engineering application value.
王传钦;李天赋;王彬;吴苓芝
东方电子股份有限公司,山东 烟台 264001东方电子股份有限公司,山东 烟台 264001山东清华康利城市照明研究设计院有限公司,山东 济南 250100烟台文化旅游职业学院,山东 烟台 264003
信息技术与安全科学
预置位自动校准纠偏特征点匹配深度学习
preset positionautomatic calibrationdeviation correctionfeature point matchingDeep Learning
《现代信息科技》 2026 (16)
31-36,6
评论