System and method for identifying trees

• US 9,508,006 B2
• Filed: 11/03/2014
• Issued: 11/29/2016
• Est. Priority Date: 11/03/2014
• Status: Expired due to Fees

• Current Assignee: Intelescope Solutions Ltd.
• Original Assignee: Intelescope Solutions Ltd.
• Inventors: Moyal, Michael, Shamir, Gil, Diamant, Erez Yaacov
• Primary Examiner(s): Yentrapati, Avinash
• Application Number:
  US14/530,852
• Publication Number:
  US 20160125222A1
• Time in Patent Office: 757 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06V 20/188 Vegetation

Abstract

A system and method of detecting trees in an image. A system and method may receive a dimension related to the trees in an input image. A two dimensional (2D) high pass filter may be applied to the input image to produce a high pass image. Objects may be marked in the high pass image based on the dimension. A processed image may be produced by associating a set of pixels in the high pass image with a respective set of grayscale values. A density operator may be applied to the processed image to identify locations with high frequency changes. Shapes may be defined to include the locations. Trees may be identified by grouping one or more shapes.

Claims

1. A method of detecting trees in an image, the method comprising:

providing an input image of an area with trees;

receiving at least one dimension related to the trees;

applying a two dimensional (2D) high pass filter to the input image to produce a high pass image, wherein the 2D high pass filter is configured to mark objects based on the at least one dimension;

producing a processed image by associating a set of pixels in the high pass image with a respective set of grayscale values, the grayscale values selected based on respective magnitudes of color variations around the set of pixels;

applying a density operator to the processed image to identify locations with high frequency changes and marking the centers of the locations;

defining shapes around the centers of the locations;

identifying canopies by grouping two or more shapes; and

identifying trees based on at least one of;

identifying trees in the high pass image by identifying clusters of marked objects and identified canopies.

2. The method of claim 1, comprising:

receiving a first dimension related to a first tree type and a second dimension related to a second tree type;

applying the two dimensional (2D) high pass filter to the input image to produce a high pass image, wherein the 2D high pass filter is configured to detect objects based on the first dimension and based on the second dimension; and

identifying, in the high pass image, trees of the first type and trees of the second type by respectively identifying clusters of objects related to the first tree type and clusters of objects related to the second tree type.

3. The method of claim 1, wherein the at least one dimension related to the trees is a size of leaves.

4. The method of claim 1, comprising:

receiving a color signature related to the trees; and

verifying the shapes correlate with trees in the input image using the color signature.

5. The method of claim 1, wherein the shapes are one of:

circles and ellipses.

6. The method of claim 1, comprising associating pixels included in the shapes with a first grayscale value and associating pixels outside the shapes with a second grayscale value.

7. The method of claim 1, comprising identifying planted areas by defining polygons, each of the polygons enclosing two or more of the shapes.

16. The method of claim 1, comprising:

receiving a first dimension related to a first tree type and a second dimension related to a second tree type;

applying the two dimensional (2D) high pass filter to the input image to produce a high pass image, wherein the 2D high pass filter is configured to detect objects based on the first dimension and based on the second dimension; and

identifying, in the high pass image, trees of the first type and trees of the second type by respectively identifying clusters of objects related to the first tree type and clusters of objects related to the second tree type.

17. The method of claim 1, comprising:

producing a processed image by associating a set of pixels in the high pass image with a respective set of grayscale values, the grayscale values selected based on respective magnitudes of color variations around the set of pixels;

applying a density operator to the processed image to identify locations with high frequency changes and marking the centers of the locations;

defining shapes around the centers of the locations; and

identifying canopies by grouping two or more shapes.

18. The method of claim 17, comprising:

receiving a color signature related to the trees; and

verifying the shapes correlate with trees in the input image using the color signature.

8. A system for detecting trees in an image, the system comprising:

a memory; and

a controller, the controller configured to;

receive an input image of an area with trees,

receive at least one dimension related to the trees,

apply a two dimensional (2D) high pass filter to the input image to produce a high pass image, wherein the 2D high pass filter is configured to detect and mark objects based on the at least one dimension,

produce a processed image by associating a set of pixels in the high pass image with a respective set of grayscale values, the grayscale values selected based on respective magnitudes of color variations around the set of pixels,

apply a density operator to the processed image to identify locations with high frequency changes and marking the centers of the locations,

define shapes around the centers of the locations,

identify canopies by grouping two or more shapes, and

identify trees based on at least one of;

identifying trees in the high pass image by identifying clusters of marked objects and identified canopies.

9. The system of claim 8, wherein the controller is configured to:

receive a first dimension related to a first tree type and a second dimension related to a second tree type;

apply the two dimensional (2D) high pass filter to the input image to produce a high pass image, wherein the 2D high pass filter is configured to detect and mark objects based on the first dimension and based on the second dimension; and

identify, in the high pass image, trees of the first type and trees of the second type by respectively identifying clusters of objects related to the first tree type and clusters of objects related to the second tree type.

10. The system of claim 8, wherein the at least one dimension related to the trees is a size of leaves.

11. The system of claim 8, wherein the controller is configured to:

receive a color signature related to the trees; and

verify the shapes correlate with trees in the input image using the color signature.

12. The system of claim 8, wherein the shapes are one of:

circles and ellipses.

13. The system of claim 8, wherein the controller is configured to associate pixels included in the shapes with a first grayscale value and associating pixels outside the shapes with a second grayscale value.

14. The system of claim 8, wherein the controller is configured to identify planted areas by defining polygons, each of the polygons enclosing two or more of the shapes.

15. A method of identifying trees in an image, the method comprising:

receiving at least one dimension related to the trees in an input image;

applying a two dimensional (2D) high pass filter to the input image to produce a high pass image, wherein the 2D high pass filter is configured to detect objects based on the at least one dimension;

producing a processed image by associating a set of pixels in the high pass image with a respective set of grayscale values, the grayscale values selected based on respective magnitudes of color variations around the set of pixels;

applying a density operator to the processed image to identify locations with high frequency changes and marking the centers of the locations;

defining shapes around the centers of the locations;

identifying canopies by grouping two or more shapes; and

identifying trees based on at least one of;

identifying trees in the high pass image by identifying clusters of detected objects and identified canopies.