Cargo boom crack diagnosis method based on Canopy-KMeans

The invention provides a Canopy-KMeans-based cargo boom crack diagnosis method, which comprises the following steps of: 1, establishing a finite element model, obtaining displacement, speed and acceleration data in x, y and z directions, and preprocessing the data; 2, performing coarse clustering, and searching a centroid set n; step 3, in the centroid set n, clustering by using a K-Means algorithm to set K = 2 to cluster whether a crack exists or not, and clustering by using K = 3 to cluster the position where the crack exists; 4, K centroids are found in the centroid set n to serve as optimal centroids; 5, training a clustering model by using the position of the optimal centroid; and step 6, evaluating a clustering result by using each index. The method can accurately determine whether cracks exist or not and the positions where the cracks exist, and has the advantages of high automation degree, short time consumption, high diagnosis accuracy and the like compared with traditional manual inspection. And the