초록 ▼

Contact-less remote-sensing crack detection and/quantification methodologies are described, which are based on three-dimensional (3D) scene reconstruction, image processing, and pattern recognition. The systems and methodologies can utilize depth perception for detecting and/or quantifying cracks. T

Contact-less remote-sensing crack detection and/quantification methodologies are described, which are based on three-dimensional (3D) scene reconstruction, image processing, and pattern recognition. The systems and methodologies can utilize depth perception for detecting and/or quantifying cracks. These methodologies can provide the ability to analyze images captured from any distance and using any focal length or resolution. This adaptive feature may be especially useful for incorporation into mobile systems, such as unmanned aerial vehicles (UAV) or mobile autonomous or semi-autonomous robotic systems such as wheel-based or track-based radio controlled robots, as utilizing such structural inspection methods onto those mobile platforms may allow inaccessible regions to be properly inspected for cracks.

대표청구항 ▼

1. A system for crack quantification, the system comprising: a storage device;a processing system connected to the storage device; anda program stored in the storage device, wherein execution of the program by the processor configures the system to perform functions, including functions to: (i) extr

1. A system for crack quantification, the system comprising: a storage device;a processing system connected to the storage device; anda program stored in the storage device, wherein execution of the program by the processor configures the system to perform functions, including functions to: (i) extract, from a binary crack map, a centerline of an extracted crack using a morphological thinning operation;(ii) determine a minimum area orientation as being equal to an orientational strip kernel having a width and a minimum correlation value from among a set of a plurality of orientational strip kernels when each is correlated with the binary crack map, wherein the minimum correlation value is the area of the crack bounded between strip edges; and(iii) for each centerline pixel, determine a measured thickness as the minimum correlation value divided by the width of the corresponding strip kernel. 2. The system of claim 1, wherein the program further includes instructions such that execution of the program by the processor configures the system to perform further functions to: (iv) determine the thickness in unit length by multiplying the measured thickness in pixels by the ratio between the working distance and the focal length. 3. The system of claim 1, wherein the program further includes instructions such that execution of the program by the processor configures the system to perform further functions to: compensate for the perspective error by aligning the view plane with the object plane. 4. The system of claim 3, wherein the perspective error is determined in accordance with the following: λx′=λxcos⁢⁢αx,where λx is the perspective-free component of the crack thickness for each centerline pixel, λx is the measured crack thickness, ax is the angle between the camera orientation vector and normal vector in the x direction of the fitted plane, and x represents either the horizontal or vertical directions. 5. The system of claim 1, wherein the instructions for (iii) include instructions to, for each centerline pixel, taking the projection of the length, in number of pixels, in a specified direction along the corresponding thickness orientation. 6. The system of claim 5, wherein the specified direction is the vertical direction or y-axis. 7. The system of claim 5, wherein the specified direction is the horizontal direction or x-axis. 8. The system of claim 1, wherein the set of orientational kernels comprise 35 kernels representing equally-incremented orientations from 0° to 175°. 9. A method of quantifying a crack, the method comprising: extracting a centerline of an extracted crack from a binary crack map using a morphological thinning operation;determining a minimum area orientation as the orientational strip kernel having a width and a minimum correlation value from among a set of orientational strip kernels when each is correlated with the binary crack map, wherein the minimum correlation value is the area of the crack bounded between strip edges; andfor each centerline pixel, determining a measured thickness as the minimum correlation value divided by the width of the corresponding strip kernel. 10. The method of claim 9, further comprising determining the thickness in unit length by multiplying the measured thickness in pixels by the ratio between the working distance and the focal length. 11. The method of claim 9, wherein for each centerline pixel, determining a measured thickness comprises taking the projection of the measured length, in number of pixels, in a specified direction along the corresponding thickness orientation. 12. The method of claim 11, wherein the specified direction is the vertical direction or y-axis. 13. The method of claim 11, wherein the specified direction is the horizontal direction or x-axis. 14. The method of claim 9, further comprising: solving a structure from motion (SfM) problem based on a plurality of images of a scene including the crack;reconstructing the three-dimensional structure of the scene; anddetermine focal length, camera center, and camera orientation. 15. The method of claim 11, further comprising using scale-invariant feature transform (SIFT) keypoints matched between pairs of images. 16. The method of claim 12, further comprising using the random sample consensus (RANSAC) algorithm is exclude outliers.