Method of feature identification and analysis
First Claim
1. A method for classifying and analyzing a digital image containing aggregate and individual features, said method comprising:
- providing a digital image;
segmenting the digital image to produce a aggregate image file, said aggregate image file comprising a first series of polygons with boundary lines;
correcting said first series of polygons by merging adjacent homogenous ones of said first series of polygons to create a larger polygon;
segmenting individual features within said larger polygons;
classifying individual features;
analyzing individual features;
wherein correcting said first series of polygons comprises unsupervised and manual correction;
wherein said manual correction comprises identifying boundary lines of first series polygon segments and merging said segments; and
wherein identifying said boundary lines comprises drawing a line or polygon to merge adjacent ones of said first series polygons.
9 Assignments
0 Petitions
Accused Products
Abstract
A method for efficiently and accurately inventorying image features such as timber, including steps of segmenting digital images into tree stands, segmenting tree stands into tree crowns, each tree crown having a tree crown area, classifying tree crowns based on species, and analyzing the tree crown classification to determine information about the individual tree crowns and aggregate tree stands. The tree crown area is used to determine physical information such as tree diameter breast height, tree stem volume and tree height. The tree crown area is also used to determine the value of timber in tree stands and parcels of land using tree stem volume and market price of timber per species.
-
Citations
39 Claims
-
1. A method for classifying and analyzing a digital image containing aggregate and individual features, said method comprising:
-
providing a digital image; segmenting the digital image to produce a aggregate image file, said aggregate image file comprising a first series of polygons with boundary lines; correcting said first series of polygons by merging adjacent homogenous ones of said first series of polygons to create a larger polygon; segmenting individual features within said larger polygons; classifying individual features; analyzing individual features; wherein correcting said first series of polygons comprises unsupervised and manual correction; wherein said manual correction comprises identifying boundary lines of first series polygon segments and merging said segments; and wherein identifying said boundary lines comprises drawing a line or polygon to merge adjacent ones of said first series polygons. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
-
-
15. A computer assisted method of timber inventory analysis, said method comprising:
-
providing a digital image of a forested area; segmenting tree stands within said digital image, said segmenting comprising performing a gradient image analysis on said digital image to identify homogenous areas, selecting seed points in the center of each homogenous area, performing seeded region growing within each homogenous area based on gradient values surrounding said seed points, correcting segmented tree stands by merging adjacent homogenous tree stands to create a larger homogenous tree stand segment, and saving resulting tree stand segments as vector files; segmenting tree crowns within said stand segment vector files, said method comprising selecting an image with a stand segment vector file overlaid, filtering the digital image to eliminate unwanted detail, selecting seed points at the center of each tree crown present within the selected stand segment vector file, performing seeded region growing within each tree crown based on pixel values surrounding said seed points, drawing boundary lines around each segmented tree crown, and saving the resulting crown boundaries as vector files; classifying tree species of said segmented tree crowns, said method comprising selecting an image with a crown vector file overlaid, manually identifying tree crowns of tree species present within the image, assigning species to the remaining unidentified tree crowns, and appending said species assignments to the crown vector file; and analyzing tree crown data, said method comprising selecting a crown vector file which includes species assignments of tree crowns, calculating the tree crown area of each tree, selecting a data model, calculating tree specific parameters based on the tree crown area and the data model, and storing resulting data in a vector file. - View Dependent Claims (16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
-
Specification