PSO-SVM (Particle Swarm Optimization-Support Vector Machine)-based classification and identification method for surface defects of bridge inhaul cable

Abstract

The invention discloses a PSO-SVM (Particle Swarm Optimization-Support Vector Machine)-based classification and identification method for the surface defects of a bridge inhaul cable. The method comprises the following steps: S1, acquiring surface image information of an inhaul cable to be detected; S2, extracting to-be-detected surface defect feature information from the to-be-detected inhaul cable surface image information; and S3, inputting the feature information of the surface defects to be detected into the PSO-SVM classifier to obtain surface defect classification and identification information of the inhaul cable to be detected. In view of the unique advantages of the support vector in solving small-sample, nonlinear and high-dimensional mode recognition, the SVM algorithm is applied to inhaul cable surface defect detection. The SVM model parameters are optimized by adopting the particle swarm optimization algorithm, so that the classification recognition rate is further improved.

Description

technical field [0001] The invention relates to the field of bridge detection, in particular to a method for classifying and identifying surface defects of bridge cables based on PSO-SVM. Background technique [0002] With the rapid development of bridge traffic construction, cable-stayed bridges and suspension bridges of long-span and super-long-span bridges are widely used. Cables are the main stress-bearing components of this type of bridge, and the reliability and durability of the cables will directly affect the safety and service life of the bridge. Since the protective layer of polyethylene (PE) or high-density polyethylene (HDPE) on the surface of the cable is exposed to the natural environment for a long time and bears alternating loads, it is prone to corrosion damage, resulting in longitudinal cracks, transverse cracks, and surface cracks on the surface of the cable. Defects such as erosion and pitted holes. These defects will seriously affect the performance of...

AI Technical Summary

Problems solved by technology

The most commonly used methods in the process of defect classification are artificial neural network identification and support vector machines, but the artificial neural network algorithm is based on the traditional statistics of asymptotic theory, and can only have ideal application effects when the number of learning samples is large.

In the case of limited samples, a well-trained artificial neural network may exhibit poor generalization ability

Images