Detection and identification algorithm of architectural defects based on image processing and computer vision

This paper focuses on the validity verification of the algorithm model of building defect detection and identification. In this paper, 2000 architectural images are collected and divided into training set and testing set according to the ratio of 7:3. Using Intel Core i7-12700K processor, Python language and Scikit-learn library to realize the algorithm, radial basis function (RBF) is selected for the support vector machine (SVM) algorithm, and the parameters (penalty parameter C=10, kernel function parameter γ = 0.1) are optimised by cross-validation. The experimental results show that the model has a high classification accuracy for different building defects, such as the accuracy of crack defect identification is 92%, The recall rate is good, and the recall rate of crack defects is 90%, The F1 value is also high, and the F1 value of crack defect is 0.91. In addition, the algorithm model has high stability and short response time. This proves the reliability and practicability of the algorithm in building defect detection, and the constructed model can effectively detect and identify various types of defects in building quality detection, which provides strong support for practical application.