Abstract:Aiming at the problems of low efficiency and poor accuracy of traditional cable insulation thickness measurement methods, an insulation thickness detection method based on adaptive local alternating genetic algorithm (ALA-GA) is proposed. The method uses the ALA-GA algorithm to search the inner and outer edges of the specimen image alternately so as to obtain the optimal insulation thickness position; the algorithm introduces the a priori structural knowledge of the specimen, and selects the initial population according to the curvature characteristics of the edges of the specimen cross-section adaptively to ensure the high quality and diversity of the genes of the initial population; puts the crossover and mutation operations in front, and locally changes the crossover and mutation methods for both the inner and outer edges of the specimen cross-section alternately, thus improving the efficiency and accuracy of the genetic method. The crossover and mutation operations are placed in the front, and the crossover and mutation modes are changed locally and adaptively for the inner and outer edges of the specimen section, so as to improve the solution speed of the genetic algorithm; in order not to lose the high-quality genes of any edge, the new population obtained after crossover and mutation and the original population perform the post-selection operation together; every time the optimal detection position is obtained, the rest of the solutions near the position are eliminated, and so on iteratively perform the ALA-GA algorithm in order to get the accurate insulation thickness detection results. Comparison experiments as well as capability verification show that the ALA-GA based method has a time cost of 0.6s~0.7s, a thinnest point measurement error of 0.0012mm~0.0015mm, an average measurement error of 0.0013mm~0.0017mm, and a measurement repeatability of 0.0018mm~0.0021mm, which are all better than the existing state-of-the-art methods, and has good generalisation capability for irregular cables.
