System and Method for Identifying Defects in a Material

Abstract

Described are computer-based methods and apparatuses, including computer program products, for identifying defects in a material. A set of features is identified based on an image of a material, wherein each feature in the set of features is a candidate portion of a defect in the material. A set of chained features is selected based on the set of features, wherein each chained feature comprises one or more features that represent candidate portions of a same defect in the material. A defect in the material is identified based on the set of chained features and the image.

What is claimed is: 
     
         1 . A computerized method for identifying a defect in a material, comprising:
 generating, by a computing device, a preprocessed image based on an original image of a material;   dividing, by the computing device, the preprocessed image into a set of sub-images;   for a first sub-image in the set of sub-images:
 determining, by the computing device, whether the first sub-image includes a feature, wherein the feature is a candidate portion of a defect in the material; and 
 if the first sub-image includes the feature, adding, by the computing device, the first sub-image to a set of feature sub-images; 
   selecting, by the computing device, a chained feature based on the set of feature sub-images, wherein the chained feature comprises one or more features that represent candidate portions of a same defect in the material; and   identifying, by the computing device, a defect in the material based on the chained feature and the original image, comprising calculating a remaining portion of the defect based on the chained feature,   wherein identifying the defect comprises:
 selecting a pair of chained features comprising the chained feature; and 
 determining if the pair of chained features satisfies a first criterion indicative of the pair of chained features being on a same defect in the material. 
   
     
     
         2 . The method of  claim 1 , wherein generating the preprocessed image comprises:
 generating a filtered image, comprising removing one or more features using a filter; and   generating the preprocessed image by subtracting the filtered image from the original image to expose the one or more features in the original image.   
     
     
         3 . The method of  claim 1 , wherein the preprocessed image is the original image. 
     
     
         4 . The method of  claim 1 , wherein the preprocessed image exposes one or more features of the material. 
     
     
         5 . The method of  claim 1 , wherein the preprocessed image comprises dark pixels and light pixels, wherein the dark pixels and the light pixels are identified based on a grey-level threshold. 
     
     
         6 . The method of  claim 5 , wherein determining whether the sub-image includes the feature comprises executing a line fitting algorithm using the light pixels, the dark pixels, or both, in the sub-image. 
     
     
         7 . The method of  claim 1 , further comprising:
 determining the feature satisfies a first set of criterions; and   determining the chained features and the calculated remaining portion of the defect satisfy a second set of criterions.   
     
     
         8 . The method of  claim 1 , wherein selecting the chained feature comprises:
 selecting a feature sub-image from the set of feature sub-images;   identifying one or more sub-images that border the selected feature sub-image, each identified sub-image including a feature; and   generating a chained feature comprising the selected feature sub-image and a sub-image from the one or more identified sub-images based on one or more constraints.   
     
     
         9 . The method of  claim 8 , wherein the one or more constraints comprises a position, an orientation, or both, of the one or more features. 
     
     
         10 . The method of  claim 1 , wherein the first criterion is based on a distance between the pair of chained features, an end direction of each feature in the pair of chained features, a turning angle of each feature in the pair of chained features, a length of each feature in the pair of chained features, or any combination thereof. 
     
     
         11 . A computer program product, tangibly embodied in a non-transitory computer readable medium, the computer program product including instructions being configured to cause a data processing apparatus to:
 generate a preprocessed image based on an original image of a material;   divide the preprocessed image into a set of sub-images;   for a first sub-image in the set of sub-images:
 determine whether the first sub-image includes a feature, wherein the feature is a candidate portion of a defect in the material; and 
 if the first sub-image includes the feature, add the first sub-image to a set of feature sub-images; 
   select a chained feature based on the set of feature sub-images, wherein the chained feature comprises one or more features that represent candidate portions of a same defect in the material; and   identify a defect in the material based on the chained feature and the original image, comprising calculating a remaining portion of the defect based on the chained feature,   wherein identifying the defect comprises:
 selecting a pair of chained features comprising the chained feature; and 
 determining if the pair of chained features satisfies a first criterion indicative of the pair of chained features being on a same defect in the material. 
   
     
     
         12 . An apparatus for identifying a defect in a material, comprising:
 a preprocessing module configured to generate a preprocessed image based on an original image of a material;   a strong feature detection module in communication with the preprocessing module configured to:
 divide the preprocessed image into a set of sub-images; and 
 for a first sub-image in the set of sub-images:
 determine whether the first sub-image includes a feature, wherein the feature is a candidate portion of a defect in the material; and 
 if the first sub-image includes the feature, add the first sub-image to a set of feature sub-images; and 
 
   a weak feature detection module in communication with the strong feature detection module configured to:   select a chained feature based on the set of feature sub-images, wherein the chained feature comprises one or more features that represent candidate portions of a same defect in the material; and   identify a defect in the material based on the chained feature and the original image, comprising calculating a remaining portion of the defect based on the chained feature,   wherein identifying the defect comprises:
 selecting a pair of chained features comprising the chained feature; and 
 determining if the pair of chained features satisfies a first criterion indicative of the pair of chained features being on a same defect in the material.