Вимірювальна система для контролювання формованої бляхи на основі трианґулювання зображень

В контексте создания методического обеспечения автоматизированных систем оптического контроля эта работа посвящается таким аспектам, как создание алгоритмических и технических решений и синтез новой процедуры измерений для обеспечения двустороннего сканирования 3D-координат поверхностей плоской детали, а также созданию новых эффективных алгоритмов обработки полученных наборов данных для вычисления пространственного распределения толщины.

В результате сканирования поверхностей детали создаётся цифровое отображение объекта дискретными наборами точек его лицевой и обратной поверхностей, для получения которого реализовано интегрирование прецизионного компьютероуправляемого вращательного стола в измерительную систему.

С помощью разработанного вычислительного алгоритма для каждой точки одной поверхности находится ближайший элемент на моделированной сеткой треугольных плоскостей противоположной поверхности и на него опускается перпендикуляр. Предложенный метод позволяет определять локальную толщину и её поверхностное распределение на критических участках детали.

Создан макет системы, получены экспериментальные данные и проверено работу разработанных алгоритмов.

Ключевые слова: измерительная система триангулирования изображений, трёхмерное сканирование, обработка изображений, контроль формированных листов металла.

Abstract

Velgan R. Image Triangulation Measuring System for Formed Sheet Metal Inspection. - Manuscript. 5000

Dissertation for Candidate of Sciences (Ph.D.) degree in Engineering on speciality 05.11.16 - Information and Measuring Systems. Karpenko Physico-Mechanical Institute of National Academy of Sciences of Ukraine. L'viv, 2006.

To ensure the quality of parts produced by sheet metal forming the inspection of products geometrical form, surface defects, thickness and possible tearing is needed. The strong demands for cost reduction, shortened development cycle and high quality involve the necessity of a fast, exact and robust measurement procedure at the stages of forming tool development, procedure and product optimisation and mass production. The availability of highly capable three-dimensional optical sensor systems based on the image triangulation principle using fringe projection method creates the possibility of fast digitising of object surface resulting in a large set of measuring points with high resolution. Existing on the market fringe projection systems enable check up of geometrical form and examination of surface structure of digitised object. To assure the stability of formed sheet metal parts, additionally to form and surface structure the thickness fluctuations in the critical zones of formed component and defects like necking, thinning (with the future risk of tear) or wrinkling should be considered.

In the context of developing automated optical inspection systems, this thesis brings two aspects into focus: enhancement of fringe projection system enabling two-side digitisation of a formed sheet metal part for creation of complete digital object model with compilation of corresponding measurement procedure and development of efficient evaluation algorithms for thickness distribution extraction from the constructed object model. This work aims at elaborating the basics for a fast, non-destructive, user independent technique for formed sheet metal part inspection able to be integrated in automated operating procedure. Achievement of this aim enabled by upgrading of fringe projection system and developing of data processing algorithms, that additionally to available geometrical form and surface inspection enables thickness distribution estimation and its visualisation over the scanned surface. For this purpose a new measurement procedure consisting of two-side object surface digitisation, pre-processing of data sets, registration of acquired point clouds in a common coordinate system, surface modelling, evaluation of local thickness and visualisation of results was elaborated.

For the acquisition of a complete object shape (including the both sides of a formed sheet metal part) an integration of motorised precision rotary table in the measuring system is realised. The integration provided by the calibration of rotary table, i.e. estimation of rotary table axis reference to coordinate system of fringe projection system using artefact consisting of two spheres at different height levels by analogy with the EN ISO 10360-3 established method for acceptance and re-verification test for coordinate measuring machines with rotary table. The part to be measured is placed on the rotary table using a suitable fixture and measured using the fringe projection system. After scanning one side, the formed metal part is rotated using a rotary table for the acquisition of the reverse side. The data sets are registered into common coordinate system by the developed registration module based on coordinate transformation and using the estimated rotation axis and angles of rotation taken from rotary table control unit.

Within this thesis a newly developed algorithm for local thickness evaluation is presented. The algorithm realises the method for estimation of local closest distances between sides of digitised formed sheet metal part. The distances are evaluated using point-to-triangle perpendiculars, where for each point of the first surface the closest linear interpolated (triangulated) surface element at the opposite side is found. An effective closest element searching procedure is realised by using of Delaunay subdivision of the acquired point clouds into set of simplexes and topology information generation. Wall thickness evaluation enables extraction and classification of defect-relevant features and its visualisation over the measured surface. The estimated area-wise thickness is visualised over the one of measured surfaces as a colour-coded map. This kind of representation enables easy localisation and documentation of the smallest wall thickness locations. Furthermore, the direct visualization helps to find reference of the wall thickness corresponding to the measured surface zone.

to prove efficient evaluation...

We realise the evaluation of the thickness as closest distance between front and back surfaces of measured part in two directions [3]. For each point of the first surface the closest distance to the linear interpolated (triangulated) opposite surface is found.

The surfaces are interpolated as tringle meshes and corresponding topology structures generated. to prove efficient evaluation...

Wall thickness evaluation enables extraction and classification of defect-relevant features and it's visualisation over the measured surface.

Keywords: image triangulation measuring system, 3D-scanning, image processing, formed sheet metal inspection.

Размещено на Allbest.ru