東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

PLANT IDENTIFICATION USING COLOR CO-...

SHEARER, SCOTT ALLAN.

Linked to FindBook

Google Book

Amazon

博客來

PLANT IDENTIFICATION USING COLOR CO-OCCURRENCE MATRICES DERIVED FROM DIGITIZED IMAGES (TEXTURE, PATTERN RECOGNITION).

Record Type:	Language materials, printed : Monograph/item
Title/Author:	PLANT IDENTIFICATION USING COLOR CO-OCCURRENCE MATRICES DERIVED FROM DIGITIZED IMAGES (TEXTURE, PATTERN RECOGNITION)./
Author:	SHEARER, SCOTT ALLAN.
Description:	190 p.
Notes:	Source: Dissertation Abstracts International, Volume: 47-07, Section: B, page: 3010.
Contained By:	Dissertation Abstracts International47-07B.
Subject:	Engineering, Agricultural. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=8625288

PLANT IDENTIFICATION USING COLOR CO-OCCURRENCE MATRICES DERIVED FROM DIGITIZED IMAGES (TEXTURE, PATTERN RECOGNITION).
SHEARER, SCOTT ALLAN.

PLANT IDENTIFICATION USING COLOR CO-OCCURRENCE MATRICES DERIVED FROM DIGITIZED IMAGES (TEXTURE, PATTERN RECOGNITION). - 190 p.

Source: Dissertation Abstracts International, Volume: 47-07, Section: B, page: 3010.

Thesis (Ph.D.)--The Ohio State University, 1986.

A method of identifying plants based on color textural characterization of canopy sections was developed. Machine tristimulus values for each pixel within an image were found using red, green, and blue filters in combination with a matrix camera. Color attributes (intensity, hue, and color) were found for each pixel from the tristimulus values. Color co-occurrence matrices were derived from the image matrices, one each for intensity, saturation, and hue. The co-occurrence matrices summarized the probability that given a pixel with an attribute level of x(,i), another pixel of attribute level x(,j) would occur as a nearest neighbor. Using the co-occurrence matrices, 11 textural features were calculated for each attribute. These features were measures of properties such as variation, correlation, and entropy. The 33 total color textural features were used in a discriminant analysis model. Distance measures between class means and a single observation were used to calculate the posterior probability of belonging to each class. An unknown observation was classified as belonging to the class with the highest probability.Subjects--Topical Terms:

1019504
Engineering, Agricultural.

PLANT IDENTIFICATION USING COLOR CO-OCCURRENCE MATRICES DERIVED FROM DIGITIZED IMAGES (TEXTURE, PATTERN RECOGNITION).
LDR:02616nam 2200253 a 45 001 935897
005 20110510
008 110510s1986 d
035 $a (UnM)AAI8625288
035 $a AAI8625288
040 $a UnM $c UnM
100 1 $a SHEARER, SCOTT ALLAN. $3 1259599
245 1 0 $a PLANT IDENTIFICATION USING COLOR CO-OCCURRENCE MATRICES DERIVED FROM DIGITIZED IMAGES (TEXTURE, PATTERN RECOGNITION).
300 $a 190 p.
500 $a Source: Dissertation Abstracts International, Volume: 47-07, Section: B, page: 3010.
502 $a Thesis (Ph.D.)--The Ohio State University, 1986.
520 $a A method of identifying plants based on color textural characterization of canopy sections was developed. Machine tristimulus values for each pixel within an image were found using red, green, and blue filters in combination with a matrix camera. Color attributes (intensity, hue, and color) were found for each pixel from the tristimulus values. Color co-occurrence matrices were derived from the image matrices, one each for intensity, saturation, and hue. The co-occurrence matrices summarized the probability that given a pixel with an attribute level of x(,i), another pixel of attribute level x(,j) would occur as a nearest neighbor. Using the co-occurrence matrices, 11 textural features were calculated for each attribute. These features were measures of properties such as variation, correlation, and entropy. The 33 total color textural features were used in a discriminant analysis model. Distance measures between class means and a single observation were used to calculate the posterior probability of belonging to each class. An unknown observation was classified as belonging to the class with the highest probability.
520 $a This method achieved an overall classification accuracy of 91% when used to discriminate between seven cultivars of containerized nursery plants. The final model used only seven textural features, four of which related to hue and the remaining three to intensity. A total of 350 observations (50 from each class) were used in the investigation. Of these, 175 were used as a training set and 175 as a test set.
520 $a This method exhibited a significant improvement over previous methods which used intensity data only. The color textural features were shown to be invariant under rotation. Classification accuracy exhibited dependency on canopy section foliage density.
590 $a School code: 0168.
650 4 $a Engineering, Agricultural. $3 1019504
690 $a 0539
710 2 0 $a The Ohio State University. $3 718944
773 0 $t Dissertation Abstracts International $g 47-07B.
790 $a 0168
791 $a Ph.D.
792 $a 1986
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=8625288