1. Institute of Electrical Engineering, Yanshan University, Qinhuangdao, Hebei 066004, China
2. Yanshan University Science Park, Qinhuangdao, Hebei 066004, China
Abstract：Based on U-net, a new method of automatic segmentation water level line is proposed and validated through a variety of scenarios. Firstly, the water and background in the original image are marked and grayscale. Then, the processed image and the original image are used to make the data set. The data set is used as the input to segment the image. Finally, all the segmented images are extracted to obtain the water level line. As the experimental results show that the automatic segmentation of U-net can accurately mark the water level and solve the influence of image background in the process of water level measurement. The recognition rate of U-net water level automatic segmentation method is above 96%, and the segmentation effect is better than other segmentation methods.
［1］Bruinink M, Chandarr A, Rudinac M, et al. Portable, automatic water level estimation using mobile phone cameras ［C］// IEEE. 14th IAPR International Conference on Machine Vision Applications. Tokyo, Japan, 2015: 426-429.
［2］Wei Y, He Y. Automatic water line detection for an USV system ［C］// IEEE. Cyber Technology in Automation, Control, and Intelligent Systems (CYBER), 2016 IEEE International Conference on. Chengdu, China, 2016: 261-266.
［3］Liu H, Jezek K C. A Complete High-Resolution Coastline of Antarctica Extracted from Orthorectified Radarsat SAR Imagery ［J］. Photogrammetric Engineering & Remote Sensing, 2004, 70(5): 605-616.
［4］Ciecholewski M. River channel segmentation in polarimetric SAR images: Watershed transform combined with average contrast maximisation ［J］. Expert Systems with Applications, 2017, 82: 196-215.
［5］Tsunashima N, Shiohara M, Sasaki S, et al. Water Level Measurement using Image Processing ［J］. Information processing society of Japan, Research report, Computer vision and image media, 2000, 121(15):111-117.
［6］赵一中, 刘文波. 基于深度信念网络的非限制性人脸识别算法研究［J］. 计量学报, 2017, 38(1): 65-68.
Zhao Y Z, Liu W B. Research on Unconstrained Face Recognition Based on DBNs Network ［J］. Acta Metrologica Sinica, 2017, 38(1): 65-68.
［7］冯家文, 张立民, 邓向阳. 基于多源融合FCN的图像分割［J/OL］. 计算机应用研究, 2018, 35(10). ［2018-04-17］. http://kns.cnki.net/kcms/detail/51.1196.TP.20171010.1732.110.html.
Fen J W, Zhan L M, Den X Y. Image segmentation based on multi-source fusion FCN ［J/OL］. Application Research of Computers, 2018, 35(10). ［2018-04-17］. http://kns.cnki.net/kcms/detail/51.1196.TP.20171010.1732.110.html.
［8］Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation ［C］// 18th International Conference on Medical Image Computing and Computer-Assisted Intervention—MICCAI 2015. Munich, Germany: Springer International Publishing, 2015: 234-241.
［9］Venhuizen F G, van Ginneken B, Liefers B, et al. Robust total retina thickness segmentation in optical coherence tomography images using convolutional neural networks ［J］. Biomedical Optics Express, 2017, 8(7): 3292-3316.
［10］李洪均, 赵志敏, 魏本征, 等. 图像噪声方差大小的测量新算法［J］. 计量学报, 2012, 33(2): 121-125.
Li H J, Zhao Z M, Wei B Z, et al. A Novel Algorithm of Image Noise Variance Measurement ［J］. Acta Metrologica Sinica, 2012, 33(2): 121-125.
［11］Salehi S S M, Erdogmus D, Gholipour A. Auto-context convolutional neural network (auto-net) for brain extraction in magnetic resonance imaging ［J］. IEEE transactions on medical imaging, 2017, 36(11): 2319-2330.
［12］Kingma D P, Ba J L. Adam: a method for stochastic optimization ［J/OL］. ［2018-04-17］. http://arxiv.org/abs/1412.6980.
2019, Vol. 40 
