We have proposed a method for identify ing water gauges on ships,aiming at the uneven brightness of images,highlights,t ilting,blurring and sticking of water gauge marks,as well as the scratches on th e hull surface and damage to water gauge marks during the recognition process.Fi rstly,we have used the improved UNet network to detect the waterline of the imag e and separate the part of ship and the part of water from the image.Secondly,we have extracted the candidate connected regions of water gauge mark through MSER in the multi-channel color space of the separated ship part image,and divided the candidate regions of water gauge mark into non-waterline candidate regions and waterline candidate regions.Thirdly,we have filtered the non-waterline cand idate regions through their BLOB and classified them with CNN.Finally,Finally,co mbined with the water gauge mark at the non-waterline that has been identified, the actual reading of the water gauge mark is calculated through the filtering a nd geometric characteristics of the connected area at the waterline.Experimental results show that the algorithm has a higher recognition accuracy rate for imag es with better quality,and the accuracy rate can reach 96.8%,and it can also hav e an accuracy rate of 80.7% for images with poor quality due to collection,ship hull and other reasons.