Аннотация:
In modern instrumentation, the number of soldered joints in printed
circuit boards can reach several thousand. Diagnostics of the soldered joints
defects within the optical wave length range is carried out using automated
diagnostic systems. A number of stages of the existing information technologies
for such systems are implemented on the basis of target functions extremum
search using the gradient estimation. In the large batches of products
production, the use of expensive automated diagnostic systems within lighting
subsystems and high cost positioning are justified. These subsystems can
provide improved noise immunity. However, in conditions when small batches
of products are produced, and at some stages (for example, when positioning by
comparison with a standard/prototype images) in general, such objective
functions can be noisy and can be multi-extremes. For such cases, information
technologies based on methods of enhanced noise immunity are required. Such
an increase in noise immunity can be provided by methods using wavelet
transformation. For this purpose, information technologies were proposed using
wavelet transformation-based procedures that improve noise immunity and
reduce the error of procedures in the diagnostic systems of printed circuit
boards and their soldered joints.
