US 12,437,531 B1
AI-based automated classification and determination of material properties of objects
Josef Johann Valentin Müllner, Pilsen (CZ); Martin Markert, Erfurt (DE); and Sebastian Johannes Brehme, Ilmenau (DE)
Assigned to e.SIC GmbH, Dornach (DE)
Filed by e.SIC GmbH, Dornach (DE)
Filed on Oct. 17, 2024, as Appl. No. 18/918,098.
Int. Cl. G06V 20/10 (2022.01); G06V 10/98 (2022.01)
CPC G06V 10/993 (2022.01) [G06V 20/194 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method for determining one or more material properties of an object, comprising:
projecting radiation of different wavelength ranges from at least one radiation source onto the object, in a temporal sequence, one wavelength range at a time, wherein the different wavelength ranges of radiation projected onto the object and the temporal sequence are pre-set by an Artificial Intelligence model engine provided by a data acquisition and evaluation system, wherein the at least one radiation source is configured to emit ultraviolet (UV) light, infrared (IR) light and visible light;
capturing, by one or more multi-spectral image acquisition devices, radiation reflected by the object and generating multiple images, and corresponding image data, according to the reflected radiation of multiple wavelength ranges, respectively, wherein a first real-time image is taken under UV light in response to the radiation source emitting UV light, wherein a second real-time image is taken under IR light in response to the radiation source emitting IR light, and wherein a third real-time image is taken under visible light in response to the radiation source emitting visible light;
forming a set of raw data to output to the data acquisition and evaluation system for data analysis, wherein the set of raw data comprises the image data;
analyzing, by the data acquisition and evaluation system, the raw data corresponding to the real-time images taken under different lights to determine one or more material properties of the object, wherein the analyzing includes, for each wavelength range emitted by the at least one radiation source and each image,
detecting one or more profiles or contoured areas and generating a corresponding raw data set;
using the corresponding raw data set to search a storage device for a data set that matches the corresponding raw data set to a preset degree, wherein the storage device stores a multitude of data sets related to sample images obtained under different lighting conditions and including material properties comprising a crack-like shape, a plant-like shape, a plant root-like shape, a stain-like shape, a chipped-off shape and a depth profile; and
outputting for each wavelength range an indication of the one or more material properties of the object in response to a match,
outputting the one or more material properties to at least one of a storage device and a display in response to the indications outputted for each wavelength range being indicative of the same one or more material properties.