Research On Construction And Application Of Pavement Image Database Based On HBase

my country’s highway management is gradually transitioning from focusing on construction to focusing on maintenance,operation and management,and gradually opens a new era of intelligent road maintenance and management.Pavement images can fully and intuitively reflect the maintenance status of highways.They are the basic data for highway information management and the core data for the analysis and evaluation of road technical conditions.They are indispensable in the field of intelligent maintenance.Due to the large amount of road image data and frequent updates and calls,it has brought huge challenges to data storage,organization,management,retrieval and service.Therefore,efficient organization and service of road image data has become the key technology in the field of intelligent maintenance information.Based on HBase big data technology,this paper researches and proposes a road image distributed database prototype system based on the characteristics of road pavement images,which realizes the efficient organization and service of road image data,supports the Liaoning highway information management platform,and realizes highway space vector data.,The seamless integration of attribute data and road image data supports the visual expression of maintenance management data and road condition evaluation information,and improves the information level and scientific decision-making capabilities of highway maintenance departments.The main research contents of the thesis are as follows:(1)Analyzed the road image data and the current highway information platform,and proposed the technical route of using the non-relational database system based on the distributed file system to realize the road image data management,and compared and analyzed the 9 mainstream distributions.After the integrated file system,the distributed database HBase based on the Hadoop parallel computing framework is adopted,and the functions such as Get,Put,Scan and Delete provided by it are used for secondary development based on the HBase API to realize the precise addition,deletion and modification of road image data.Check function.(2)The HBase distributed database cluster was built.Adopting the Zookeeper+Hadoop+HBase+JDK technology route,researched and proposed HBase distributed database optimization strategy,conducted experiments on the KeyValue value that has the greatest impact on HBase read and write performance,and determined the optimal value range of throughput,latency and average running time;Research the theory and method of secondary index,put forward a hash index algorithm based on hash index algorithm,reverse order algorithm and RowKey hashing algorithm,combined with database tables to form a column secondary index,and improve the read performance of a single piece of data to millisecond level.(3)Based on the Spring Boot framework,HBase API and related Java code writing,the automatic reading and writing microservice interface of road image is realized,and it provides support for batch storage,batch reading and data management of image data.The results are finally applied to "Liaoning Province".Good results have been achieved in the “ Highway Big Data Basic Management Platform” and “Pavement Disease Detection Cloud Platform”.