应用图像处理技术的舱室液位识别算法研究

Research on Cabin Liquid Level Recognition Algorithm Using Image Processing Technology

为了快速准确地定位船舶运行过程中的损害隐患,改善船舶舱室中液位识别自动化程度低的缺点,基于图像处理技术,利用滤波、灰度转化等方法完成对采集图像的预处理,通过直方图均衡化及图像分割技术实现对图像的进一步处理,在对处理后的图像进行形态学操作及轮廓逼近后,提取出液位检测结果,将水箱模型中的测量结果与实际值进行比较,得出相对误差。结果表明,应用图像处理技术的舱室液位识别算法较传统液位识别算法精度较高,相对误差达3.27%。该算法的液位检测识别准确率为96.73%,满足船舶实际运行需求,且由于该算法可封装成智能识别系统,具有适应能力强、部署灵活的特点,具备向船舶实际作战中损害管制系统扩展的应用空间。

In order to locate the damage during the operation of the ship quickly and accurately,and solve the automation problem of liquid level recognition in the ship's cabin,this article tries to use filtering,gray scale conversion and other methods based on image pro⁃cessing technology to complete the preprocessing of the collected image.Afterwards,this article continue to do a further processing of the image through histogram equalization and image segmentation technology.When the contour is approximated and the liquid level detection result is extracted,the measurement result in the water tank model is compared with the actual value to obtain the relative er⁃ror.The experimental result shows that the cabin liquid level recognition algorithm using image processing technology is more accurate than the traditional liquid level recognition algorithm,reaching 3.27%in relative error.The accuracy of this algorithm has reached 96.73%,which meets the actual operational requirements of ships.Because the algorithm can be packaged into an intelligent identifica⁃tion system,it has the characteristics of strong adaptability and flexible deployment,which can be used for damage control system de⁃ployed in actual ships.