Optical method and apparatus for identifying wood species of a raw wooden log

1. An optical method for identifying wood species of a raw wooden log, comprising the steps of:

i) directing light onto at least a portion of a peripheral surface of said raw wooden log, said illuminated portion presenting light reflection characteristics being substantially representative of said log peripheral surface;

ii) sensing light reflected on the illuminated representative log portion to generate reflection intensity image data associated with said log peripheral surface, said reflection intensity image data including color image data;

iii) subdividing said reflection intensity image data into a plurality of image data regions each containing a preset number of image pixels;

iv) analyzing each of said image data regions to generate associated texture data;

v) analyzing the color and texture data associated with each of said image data regions to assign to each thereof a probable one of a plurality of species indications; and

vi) selecting a majority one of said assigned species indications as said wood species identification of the raw wooden log.

2. The method according to claim 1 , wherein said wood species identification is one of species of spruce and fir.

3. The method according to claim 1 , wherein said analyzing step v) is performed with a classification model previously trained with a set of raw wooden logs representative of said plurality of species indications.

4. The method according to claim 2 , wherein said classification model is a neural network model.

5. The method according to claim 1 , wherein color image data is defined in one of a LAB color space and OHTA color space.

6. The method according to claim 5 , wherein said color image data is expressed as values selected from the group consisting of mean values, standard deviation values, variance values, or any combination thereof.

7. The method according to claim 1 , wherein said analyzing step iv) includes:

a) determining local binary patterns for each of said image data regions;

b) calculating a histogram of said local binary patterns to generate said associated texture data.

8. The method according to claim 1 , wherein said selecting step vi) includes calculating a histogram of said assigned species indications.

9. An optical apparatus for identifying wood species of a raw wooden log, comprising:

an optical sensor unit including:

a light source configured for directing light onto at least a portion of a peripheral surface of said raw wooden log, said illuminated portion presenting light reflection characteristics being substantially representative of said log peripheral surface; and

an imaging sensor having a sensing field oriented to capture light reflected on the illuminated representative log portion and being configured to generate reflection intensity image data associated with said log peripheral surface, said reflection intensity image data including color image data; and

data processing means programmed for subdividing said reflection intensity image data into a plurality of image data regions each containing a preset number of image pixels, analyzing each of said image data regions to generate associated texture data, analyzing the color and texture data associated with each of said image data regions to assign to each thereof a probable one of a plurality of species indications, and selecting a majority one of said assigned species indications as said wood species identification of the raw wooden log.

10. The apparatus according to claim 9 , wherein said wood species identification is one of species of spruce and fir.

11. The apparatus according to claim 9 , wherein said data processing means is programmed for analyzing the color and texture data with a classification model previously trained with a set of raw wooden logs representative of said plurality of species indications.

12. The apparatus according to claim 11 , wherein said classification model is a neural network model.

13. The apparatus according to claim 9 , wherein color image data is defined in one of a LAB color space and OHTA color space.

14. The apparatus according to claim 13 , wherein said color image data is expressed as values selected from the group consisting of mean values, standard deviation values, variance values, or any combination thereof.

15. The apparatus according to claim 9 , wherein said data processing means is programmed for processing each of said image data regions to generate associated texture data through a determination of local binary patterns for each of said image data regions, followed by calculation of a histogram of said local binary patterns.

16. The apparatus according to claim 9 , wherein said data processing means is programmed for selecting a majority one of said assigned species indications as said wood species identification of the raw wooden log through a calculation of a histogram of said assigned species indications.

17. An optical apparatus for identifying wood species of a raw wooden log, comprising:

a first optical sensor unit including:

a first light source configured for directing light onto at least a portion of a peripheral surface of said raw wooden log, said illuminated portion presenting light reflection characteristics being substantially representative of said log peripheral surface; and

a first imaging sensor having a sensing field oriented to capture light reflected on the illuminated representative log portion and being configured to generate color image data;

a second optical sensor unit including:

a laser source configured for directing a linear-shaped laser beam onto the portion of the peripheral surface of said raw wooden log to form a reflected laser line onto said log peripheral surface;

a second imaging sensor having a sensing field oriented to capture a two-dimensional image of said reflected laser line to generate corresponding two-dimensional image data, wherein said linear-shaped laser beam is directed at an angle with said sensing field; and

first data processing means programmed for deriving profile-related image data from said corresponding two-dimensional image data; and

second data processing means programmed for subdividing said color image data and profile-related image data into a plurality of image data regions each containing a preset number of image pixels, analyzing each of said profile-related image data regions to generate associated texture data, analyzing the color and texture data associated with each of said image data regions to assign to each thereof a probable one of a plurality of species indications, and selecting a majority one of said assigned species indications as said wood species identification of the raw wooden log.