Diseño y construcción de un escáner bifocal para la reconstrucción visual de objetos utilizando procesamiento de imágenes mediante filtrado en labview y matlab

Abstract

Three-dimensional reconstruction of objects and images in this set in a wide field of technological development of artificial vision, this three-dimensional reconstruction allows the user to digitally images of objects in the PC that costs are high or antiques, to study the object without manipulation was designed and a three-dimensional bifocal scanner objects , for which we used was built: two webcams, a laser, a swivel base, a main computer which enables controlling the movement of the swivel base and is where captured by the two webcams, that are subsequently analyzed and processed images, whereby the projection points of the laser line over the object coordinate x is obtained is stored, and the units of the coordinates obtained are given in pixels, the z - coordinate is determined by depth mathematical calculations using a computer vision algorithm, and once obtained the coordinates x, y, z is not necessary to connect all the points in a single three-dimensional point cloud. The team developed in this project aims to scan objects up to 20x15 cm, following the steps of processing, edge detection, extraction of coordinates by laser light is projected onto the object, and once the above process completed is obtained three-dimensional coordinates of the scanned object to reconstruct the 3D image of the object. Consequently the coordinates x, y, z of the images obtained by the pair of webcams are compared with the coordinates x, y, z of the Kinect sensor to obtain a percentage error of 1.789 % in the calculation of their actual coordinates of the image and 98.20% accuracy.