Key Techniques Research On3D Vision Surface Measurement System Based On Linear Structured Light

As the3D vision treasurement technology was widely used in industrial ares、aerospaceares and medical ares and so on,more and more scholars focus on it.Through the research,lotsof excellent measuring techniques have been producted.In these techniques,the measurementtechnique based on linear structured light has been more widely used because of itsadvantages of non-contact and real-time and high percise.In3D vision surface measurement system based on Linear Structured Light, First,thelinear structured light projected onto the object surface,then the light stripe was produced onsurface.the light stripe would be deformation because of modulation of the object surfacetopography. The deformation light was collected with a CCD camera. Then we had cameracalibration. The location of light stripe in the two-dimensional image could determine thecorresponding spatial coordinates on object surface through calculation of the processor.Finally,the three-dimensional topograghy on the surface of the object tested had been obtained.Camera calibration is a critical step in the process. This article adopts the classic checkboardcalibration method invented by zhangzhengyou.Extracting central positions of light stripe plays a very important role in3D visionsurface measurement system based on Linear Structured Light.In this paper,a new kind ofcentral positions extraction algorithm of light stripe has been proposed on the basis of theexisting extraction algorithms. Before the central extraction of light stripe,First of all,thecollected image must be pretreated.In new algorithm,The median filtering method wasadopted.Otsu algorithm was used to carry out threshold segmentation.then by using thethresholding method to gain crude extractive center line of the light stripe. the approximatenormal directions of the light stripe were acquired by using the differential mean valuetheorem.Then, in the normal directions, the central positions of the stripe were gained with thegray barycenter method.The new algorithm is superior to Hession Matix extraction method interms of complexity,It also has faster extraction rate、higher extraction precision and strongpracticability.