Semi-automatic extraction of linear features from image data including path material type attribution

Abstract

Method for editing a vector set associated with an extracted linear feature in a remotely sensed image, the vector set defining a path and being tied to a geographical location. The method includes displaying the path in a graphical display. Once the user activates a smart editing tool, the user establishes a region of influence centered around a cursor. The region of influence is configured to respond to cursor movements. The user specifies a point near the path and moves the cursor to it, bringing the region of influence along. Any error in the vector set of the path is automatically corrected in real time using image-based logic. The user then previews the correction on the graphical display and implements it, updating the path. The updated path is displayed in real time in the graphical display.

What is claimed is: 
     
         1 . A computer-implemented method for attributing a characteristic to an extracted linear feature in a remotely sensed image, the method comprising:
 providing a vector set defining an elongated path in an image, the elongated path being longer than it is wide, and being tied to a geographical location;   displaying the path in the image in a graphical display; and   via a computer, attributing a path material type to the path.   
     
     
         2 . A computer-implemented method as defined in  claim 1 , wherein the path material type is determined automatically by the computer. 
     
     
         3 . A computer-implemented method as defined in  claim 2 , wherein the path material type is determined automatically by using image-based logic. 
     
     
         4 . A computer-implemented method as defined in  claim 3 , wherein the image-based logic includes a Maximum Likelihood algorithm. 
     
     
         5 . A computer-implemented method as defined in  claim 4 , wherein the algorithm is based on a probabilistic mixture model, wherein each mixture component is a probability distribution for a particular path material type. 
     
     
         6 . A computer-implemented method as defined in  claim 1 , further including providing a computer-human interface through which the path material type is manually assigned by a human operator. 
     
     
         7 . A computer-implemented method as defined in  claim 1 , wherein the attributed path material type is stored in a vector file associated with the path. 
     
     
         8 . A computer-implemented method as defined in  claim 1 , wherein the path is a road.