ARTIFICIAL INTELLIGENCE ASSISTED JUDGMENT
Artificial intelligence is a simulation of the information process of human consciousness and thinking. Its research includes robotics, language recognition, image recognition, natural language processing and expert systems. In recent years, artificial intelligence has developed rapidly in various fields and will profoundly and thoroughly change society. Of course, there is no room for development in the field of engineering inspection.
At present, in the field of engineering inspection, there are the following problems that need to be solved.
◆ Over-reliance on the technical level and experience of the inspectors
◆ Due to the limitations of human senses, many parameters cannot be rationally utilized.
◆ Historical data, experience accumulation is slow, and there are ups and downs
◆ The threshold of defect determination has a large impact and there are human factors.
◆ Lack of organic links between different agencies and different people
In order to solve the above-mentioned industry pain points, we are developing an auxiliary detection and determination system based on AI.
Model and accuracy
The model mainly adopts Bayesian network (BN), neural network (NNS) and set judgment; the defect identification for indoor concrete test block: the comprehensive identification accuracy is above 95%; for the dozens of prestressed tunnel grouts we collect The comprehensive recognition accuracy of the empirical engineering of compactness is 88~98%.
◆ All resolution processes can be run automatically on the database server
◆ Significantly reduce the intensity and difficulty of the tester, saving analysis time
◆ Guaranteed detection accuracy and continuous improvement
◆ In the near future, the prestressed tunnel grouting and the prefabricated structure grouting test will be gradually promoted in the future.
◆ Partner: Recently recruited partners to participate in system testing and model training;
◆ Customer: It is expected to officially provide services to customers in September this year.