By Xiang Linhao, CCS Science and Technology Innovation Center

The hull structure survey of ships in service generally includes the overall survey, close–up survey and thickness measurement, to evaluate the overall condition of the hull  structure, coating condition,structural defects and corrosion conditions.Traditional hull structure surveys have disadvantages such as the high cost, low efficiency and high risk . After he survey, the survey records of the hull structure status are generally kept in the  form of records, reports and others, which are poorly readable . In order to solve these problems,  China Classification Society (CCS) has carried out digital survey of the  hull structure of ships in service from three aspects:  "survey data acquisition means,intelligent diagnosis method, and 3D digital model" .  Specifically, the hull structure survey images and thickness measurement data are collected by drones or robots, to improve the data acquisition efficiency;the image  recognition technology is used to intelligently identify the key structural areas of the hull structure and the typical hull structure defects, with the defect identification accuracy improved; the survey  data collected by drones or robots is associated with the 3D digital model of the hull structure, and the status of the hull  structure is evaluated based on the 3D model, to assess the coating condition,corrosion and structural grade comprehensively and accurately .

I. Data acquisition for digital survey of hull structure

In recent years, CCS has applied drones in the close-up survey of the hull structure of more  than  30  large  oil  tankers  and bulk carriers, and promoted the technical achievements to the industry, leading the practical  application  of large  domestic shipping companies and shipyards . Through the acquisition of hull structure images and thickness measurement data by drones or robots, the popularization and application of RIT can be effectively promoted, the survey efficiency can be improved, and the survey costs can be reduced, so as to improve the digitalization level of survey, and provide decision-making support for the surveyor to accurately evaluate the  status of hull structures and for shipping companies to formulate repair plans .

The  drones  developed by  CCS  and interested parties in the industry have the following technical characteristics: 1. They are  well  adapted to the  enclosed  space of the ship, and have the capabilities of autonomous navigation, stable flight and hovering, intelligent perception of obstacles and  autonomous  obstacle  avoidance.2.  They  support  the  high - definition image  acquisition  and  electromagnetic ultrasonic  thickness  measurement,  and the pan -tilt  system is equipped with  an image  acquisition unit  and  a multi - axis manipulator, which is flexible and capable of approaching and even directly contacting the hull structure to collect the images and thickness data . 3. They run in a wired or wireless mode .

In view of the scenario characteristics and  survey  requirements  of the  close -up survey and thickness measurement of the hull  structure, the robots developed by CCS and the interested parties of the industry  have  the  following  technical characteristics:  1.  They  are  flexible and  compact,  light  in  weight,  adaptive to  the  narrow  spaces,  and  capable  of continuously climbing over the structural obstacles .  2.  They  have  the  capabilities for  high - definition  image  acquisition and electromagnetic ultrasonic thickness measurement,  so that the  "back" of the structure can be observed easily . 3. They run in a wired or wireless mode .

In the cargo tanks and ballast water tanks of floating production storage and offloading tankers (FPSO), 300,000 DWT very-large crude oil tankers and  180,000 DWT large bulk carriers, CCS has applied the unmanned aerial vehicles and robots jointly developed with industrial units to the close-up survey and thickness measurement of the  hull  structure,  and  carried  out the  shipboard test and verification . The capabilities and limitations of UAVs and robots have been fully tested in terms of their adaptability to the cabin environment, t he  ability  t o  approach  or  climb  over structures,  survey  data  acquisition and  processing  capabilities,  thickness measurement  accuracy  and  equipment safety performance,  which  is  a  crucial step for their application in full-scale ship survey .

II. Intelligent diagnosis of hull structural defects based on image recognition technology

Image recognition technology is to automatically identify and understand the digital image contents through computer algorithms and machine learning methods . The  technology  can  be  applied  to  the close-up survey of the hull structure . With this technology, the defects and damages on the  surface of the hull  structure can be automatically detected and identified through the analysis of the digital image,to realize the intelligent detection,  andimprove  the  detection  efficiency  and accuracy .

According to the definitions of “Good”,“Fair” and “Poor” on coating status in the Guidelines for Maintenance  and Repair of Oil Tanker Ballast Tanks and Liquid Cargo/Ballast Tanks of the International Association of Classification  Societies,the coating condition has been evaluated,mainly  depending  on  the  location,  area and degree of coating peeling or rusting .With reference to the requirements of the current mainstream image  segmentation algorithm based on deep learning for the data set, different types of corrosion data sets have been established, and the sample data has been cleared and enhanced . Then,a corrosion detection model of the hull structure has been built using the semantic segmentation method for coating condition detection .

In view of the characteristics of large scale variations of hull  structure cracks,many  small  cracks,  and  small  sample size, CCS has built a hull structure crack detection  model  based  on  the  target detection algorithm .

In  addition, CCS has developed  an intelligent diagnosis  system for the hull structural  defects  based  on  the  image recognition technology, allowing for the identification of the key structural areas such as large bracket toe ends, longitudinal bone penetrations, and grooved bulkheads, as well as the coating damage, cracks and other typical hull structural defects .

III. 3D model-based assessment of hull structural condition

CCS  has  developed  an  auxiliary  decision - making  system  for  the  hull structure assessment, which automatically receives  the relevant  information  such as  hull  structure  images  and  thickness measurement  data  collected  by  UAVs and  robots,  and  evaluates  the  coating conditions,  corrosion  and  structural conditions based on the image recognition results, thickness measurement data and 3D models of hull structures . Meanwhile, the decision-making system also assesses the  grade  of ship  cabins  and  structural units, visually displays the hull structural condition  and  rating  results  (coating, defects  and  corrosion),  and  estimates the  amount  of repair  work  for the hull structure (including the painting area and the structural repair quantity) .

In addition, CCS has established the rules for data transmission interface with the  service  providers  related  to  UAVs and  robots,  automatically  associating the hull  structure images  and thickness measurement  data  collected  by  UAVs or robots with the 3D model of the hull structure  by  structural  units  through receiving and automatically parsing the standard data packets sent, including the survey tasks,  survey  images, thickness measurement data and other information related to structures .

Combining the image technology and the 3D model of the hull structure, CCS has established the rules for identifying the coating status of the hull structure and the grade of the structure status, allowing to automatically evaluate the coating and structure  status based  on the results  of key  structural  areas,  structural  cracks,and coating damage of the hull structure identified by image recognition technology .With the thickness measurement data and the  3D model of the hull  structure, CCS has established the rules for identifying the corrosion condition of the hull structure,allowing  to  automatically  evaluate  the corrosion condition of the hull structure based  on  the  analysis  of the  thickness measurement  data . Moreover,  CCS has established the rules identifying for cabin conditions based on the above methods,allowing to automatically evaluate the cabin status .

Based on the intelligent recognition of the  survey  image  and the  automatic analysis of the thickness measurement data of the hull structure, and combined with the 3D model of the hull structure, CCS has established the rules for repair amount estimation of the structural unit . Thus, the coating repair area of the structural unit due to the damage of the coating, and the replacement weight of structural units due to  corrosion,  as  well  as the replacement weight of structural units due to structural defects, can be calculated .