Date of Award
The goal of this project was to create a software tool to unwrap digital images of cylindrical rock cores and also to create a tool to measure the accuracy of the unwrapping transform.
Measurements of an object can be taken directly from an image if the object is planar and the image plane of the camera is perpendicular to the object. If these conditions are met then picture scale allows a user to relate measurements in the image to a real world coordinate system after correcting for radial distortion. But if the object in the image is not planar or is not perpendicular to the image plane then additional corrections or geometric transformations must be applied to the object in the image.
The project is focused on creating a coded software solution, using digital images from an off the shelf-camera, to transform or unwrap single images of cylindrical rock cores so that all objects in the image are co-planar. In addition to tools created to unwrap digital images a second software tool was created to verify the accuracy of the unwrapping software.
The project is broken down into three major components. The first is a technique for taking pictures of rock core, or image acquisition. A single rock core is wrapped with a paper grid pattern that is used to determine a pixel scale for a series of images of a rock core sample. The grid wrapped core is placed in a carriage made of a set of two rollers. This carriage is then centered below a camera that is remotely controlled from a desktop computer. Two halogen lights are used to evenly illuminate the surface of the core and an image is taken. Once the grid pattern is photographed the pattern is removed and a series of images of the outer surface of the core are taken.
The next phase of the unwrapping is to take the original image and pass it through a series of software modules that will assist the user to perform a rectification of the original images. First the scale of each pixel in the image is found using images of the original grid pattern wrapped around the core. Next the core itself is isolated from other background objects present in each image. Then a geometric transform is applied to the image of the core, using the unique pixel scale, with the goal to unwrap each image so that all objects in the image are co-planar and no longer suffer any distortion.
In order to measure the effectiveness of this transform on rectifying all areas of the image, a second software tool was created. This software tool allows the user the ability to measure the spacing of gridlines in the original image of the grid wrapped core and compare these measurements to the same image after it has been unwrapped. The code returns a database of measurements for every gridline present in a image as well as different plots of the data for the user to make final conclusions as to the effectiveness of unwrapping on each area of the core images.
Zosel, Aaron, "Development and Application of a Method for Unwrapping Single Images of Cylindrical Objects" (2015). Graduate Theses & Non-Theses. 24.