In the assembly of aircraft, the measurement methods for determining the correct sizing of shims is very important to assure correct structural loading. With the advent of advanced composite structures, additional emphasis on assembly verification and its automation is very important. The measurement of the gap between layers of structural materials that are fastened together is of prime interest. A prototype inspection system using a boroscope and machine vision hardware is discussed. Special hardware probes and lighting considerations are described. Images of the gaps have been taken with and without shims present that will be typical of future aircraft fabrication. The dimensional measurement method is done with the aid of the human eye and machine vision algorithms. Automation is necessary to decrease the inspection time required and to provide automatic documentation of the process. The data that is taken can subsequently be fed to an automatic shim fabrication machine for total process automation. The comparison between conventional machine vision metrology measurement methods and neural network software methods are presented. With the images grabbed with a frame grabber, contrast threshold techniques and equalization methods are used for image enhancement. Edge finding methods are presented for finding and measuring the gaps in the assembly. Limitations for conventional machine vision metrology measurement method algorithms are discussed, and the usage of a neural network to solve these problems is presented. The models of the neural network are discussed and the testing results from the images shown.
展开▼
