Accuracy Assessment of Industrial Heritage Mapping Based on GA+BP Neural Networks Forecast

With the gradual implementation of the protection project today, when the heritage has more favorable preservation conditions, the development of The Times requires us to change the original protection idea on the basis of “repair”, and turn the focus to science, systematic research, and complete management of heritage information. Starting from the collection and collation of heritage information, a set of relatively perfect heritage archives database has been established. It has promoted the historic transformation from passive rescue protection to active systematic and preventive protection. The establishment of a perfect heritage information database is the premise to achieve this transformation, and the establishment of an efficient and practical data management system is the foundation. This puts forward high requirements for the record and management of heritage. Therefore, this paper proposes to use GA+BPNN to predict the surveying and mapping accuracy of industrial building heritage, pay attention to the processing of “information”, and carry out preliminary processing of surveying and mapping information according to the “ideal model”. After experimental training on GA+BPNN model, the accuracy forecast of industrial building heritage through BP neural channel is made. Through comparative analysis, GA+BPNN has a better effect on the accuracy forecast of industrial architectural heritage. There are absolute errors in the GA+BP combination model, and the relative errors and predicted values are 0.7488mm, 0.0972% and 13.0772 mm, respectively. The forecast results are good. It also has certain reference significance for the surveying and mapping of industrial building heritage.