https://orcid.org/0009-0000-0837-4806
ABSTRACT
With the rapid development of science and technology, unmanned aerial vehicle inspection technology has also been widely applied in various fields of society. To solve the problem of low efficiency in current long-distance water channel inspection and slope damage diagnosis methods, this study proposes a long-distance water channel safety inspection method that combines building information models and drones. It combines a support vector machine, genetic algorithm, and principal component analysis algorithm to construct a slope failure diagnosis (SFD) model based on this algorithm. In the performance comparison, it was found that the average accuracy and average runtime were 98.18% and 0.19 s, which were superior to the compared algorithms. Subsequently, the effectiveness analysis of the detection method showed that it was effective. Finally, in the performance comparison of the SFD model, the F1 values for the four types of slope failures were 98.4, 96.3, 97.8, and 96.9%, all of which performed the best. This indicates that the proposed long-distance water transmission channel inspection and diagnosis methods have good performance and practicality, and can provide a theoretical basis for water transmission channel inspection and disaster diagnosis.
HIGHLIGHTS
This study proposes a long-distance water channel safety inspection method that combines building information models and drones.
It combines a support vector machine, genetic algorithm, and principal component analysis algorithm to construct a slope failure diagnosis model based on this algorithm.
The proposed methods have good performance and practicality.
