| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Horticulture, Univ. of Arkansas, 308 Plant Sci. Building, Fayetteville, AR 72701
* Corresponding author (karcher{at}uark.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: AS, ammonium sulfate • DAT, days after treatment • DGCI, dark green color index • HSB, hue, saturation, and brightness • NTEP, National Turfgrass Evaluation Program • PCU, polymer-coated urea • RGB, red, green, and blue • SCU, sulfur-coated urea
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several techniques have been used to objectively measure turf color, including reflectance measurements (Birth and McVey, 1968), chlorophyll and amino acid analysis (Johnson, 1973; Nelson and Sosulski, 1984), and comparison with standardized colors (Beard, 1973). All of these methods have certain disadvantages compared with subjective color ratings. Reflectance, chlorophyll, and amino acid measurements all require relatively expensive equipment, and transport of samples to a laboratory for analysis. In addition, correlations between color and chlorophyll or amino acid measurements are either species or cultivar dependent. The use of standardized charts to measure turf color is effective, but results in qualitative descriptions of color that are not possible to statistically analyze with traditional ANOVA techniques.
Recently, Landschoot and Mancino (2000) demonstrated that the color of creeping bentgrass cultivars could be successfully quantified with a colorimeter. Values from the colorimeter were significantly correlated with visual color assessments averaged across five evaluators. Other researchers have successfully used colorimeters to evaluate varying turf color due to seasonal changes (Kimura et al., 1989) or differences among cultivars and genetic lines (Thorogood et al., 1993). Although promising, a potential shortcoming of the colorimeter used in those studies is the relatively small measurement area (<20 cm2). In the absence of extremely uniform surface conditions, numerous subsample measurements with the colorimeter would be necessary to accurately represent the color of typical turfgrass field plots.
In recent years, digital photography has become a common and affordable means for the scientific community to document and present images. Digital cameras, in conjunction with image analysis software, are being used to quantify wheat (Triticum aestivum L.) senescence (Adamsen et al., 1999) and canopy coverage in wheat (Lukina et al., 1999) and soybeans [Glycine max L. (Merr.)] (Purcell, 2000). Recently, digital image analysis was used to quantify turf coverage with increased precision over more traditional evaluation methods (Richardson et al., 2001). Through digital photography, researchers can instantaneously obtain millions of bits of information on a relatively large turfgrass canopy. For example, a digital image taken of a turf plot using a 1280 x 960 pixel resolution contains 1 228 800 pixels, with each pixel containing independent color information about the turf plot. Therefore, digital photography and subsequent image analysis may be capable of quantifying turfgrass color in field experiments.
The information contained in a digital image includes the amount of red, green, and blue (RGB) light emitted for each pixel in the image. Although it may be intuitive to use the green levels of the RGB information to quantify the green color of an image, the intensity of red and blue will confound how green an image appears. To ease the interpretation of digital color data, RGB values can be converted directly to HSB values that are based on human perception of color (Fig. 1). In HSB color description, hue is defined as an angle on a continuous circular scale from 0 to 360° (0° = red, 60° = yellow, 120° = green, 180° = cyan, 240° = blue, 300° = magenta), saturation is the purity of the color from 0% (gray) to 100% (fully saturated color), and brightness is the relative lightness or darkness of the color from 0% (black) to 100% (white) (Adobe Systems, 2002). Among HSB, hue has been found to be the best indicator of the visual color of a turf (Landschoot and Mancino, 2000; Thorogood et al., 1993). However, preliminary work at the University of Arkansas has demonstrated that visual differences in turf color were sometimes the result of color saturation differences between turf plots rather than hue differences (Karcher, 2000, unpublished data).
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
260 kilobytes per image). Camera settings included a shutter speed of 1/400 s, an aperture of f/4.0, and a focal length of 32 mm. Images were downloaded to a personal computer for subsequent analysis. The average RGB levels of the digital images were calculated using SigmaScan Pro version 5.0 software (SPSS, 1998). The entire image was selected for analysis by including all possible hue and saturation levels in the color threshold option of the software. The average red, average green, and average blue measurement settings were used to obtain the average RGB levels for an image. The average RGB levels were then pasted into an MS Excel spreadsheet (Microsoft Corporation, 1999) created by the authors to automate the conversion of RGB to HSB values. The programmed formulas in the spreadsheet converted absolute RGB levels (measured on a scale of 0 to 255) to percentage RGB levels by dividing each level by 255. Percentage RGB levels were then converted to average HSB levels by the following algorithms (Adobe Systems, 2002):
Hue
 |
 |
 |
Saturation
 |
Brightness
 |
Camera Calibration
A series of digital images were taken of color chips from Munsell Color Charts for Plant Tissues (GretagMacbeth LLC, New Windsor, NY). Six images of varying hue were collected, ranging from yellowish green to green (chip numbers 5Y 6/6, 2.5GY 6/6, 5GY 6/6, 7.5GY 6/6, 2.5G 6/6, 5G 6/6). Eight images of varying saturation were collected, ranging from grayish green to bright green (chip numbers 7.5GY 6/2, 7.5GY 5/2, 7.5GY 6/4, 7.5GY 5/4, 7.5GY 6/6, 7.5GY 5/6, 7.5GY 6/8, 7.5GY 5/8). Ten images of varying brightness were collected, ranging from light green to dark green (chip numbers 7.5GY 8/4, 7.5GY 7/4, 7.5GY 6/4, 7.5GY 5/4, 7.5GY 4/4, 7.5GY 8/6, 7.5GY 7/6, 7.5GY 6/6, 7.5GY 5/6, 7.5GY 4/6). These Munsell color chips were chosen because they covered a relatively broad range of HSB levels and visually corresponded with plant tissue HSB levels typical of turfgrass (Beard, 1973). Calibration images were taken under dark conditions using only the camera flash as a light source. The images were analyzed for HSB levels using the methods described above. To determine the accuracy of HSB measurement with digital image analysis, the actual HSB levels of the Munsell color chips were determined using Munsell Conversion software version 4.1 (Munsell Color, 2000).
Three separate linear regression analyses were performed using PROC REG in SAS Statistical Software (SAS Institute., 1996). The H, S, and B values from digital image analysis were analyzed as the independent variables and the actual H, S, and B values of the Munsell color chips were the dependent variables. For each HSB parameter, digital image analysis was considered to significantly detect color differences among color chips when the slope of the regression line was significantly different from zero (P < 0.05) (Freund and Wilson, 1993).
Nitrogen Fertility Color Differences
Two ongoing N fertility field studies were used to assess the ability of digital image analysis to quantify visual color differences among turf plots due to N treatments. The first experimental area was established with ‘Meyer’ zoysiagrass during the summer of 1996 on a silt loam (Typic Hapludult, pH 6.2). Individual plots were 1.4 m2 and mowed at a height of 1.9 cm. The second experimental area was a ‘Crenshaw’ creeping bentgrass putting green built in 1998 according to USGA recommendations (United States Golf Association, 1993). Individual plots were 1.5 m2 and mowed at a height of 0.4 cm. Both experimental areas were located at the University of Arkansas Research and Extension Center in Fayetteville, AR.
The zoysiagrass study consisted of two treatment factors, N source (7 levels) and N rate (3 levels). The N source treatment levels included: (i) 100% ammonium sulfate (AS); (ii) 100% polymer-coated urea (PCU); (iii) 100% sulfur-coated urea (SCU); (iv) 33% AS, 67% PCU; (v) 33% AS, 67% SCU; (vi) 67% AS, 33% PCU; and (vii) 67% AS, 33% SCU. Each N source was applied at three N rate levels: (i) 4.8, (ii) 7.2, and (iii) 9.6 g m-2. Each of the resultant 21 fertility treatments was replicated four times in a randomized complete block design. Treatment applications were made in mid-May and mid-August in 2000.
The creeping bentgrass study consisted of one treatment factor, N rate (7 levels). The N rate treatment levels included 0, 1, 2, 3, 4, 5, and 6 g m-2. The N source for all treatments was methylene urea. Each N rate was applied four times in a completely randomized design. Treatment applications were made monthly from June through September in 2000.
Digital images were collected from each plot on 28 Sept. 2000 on the zoysiagrass study [44 d after treatment (DAT)] and on 16 Nov. 2001 on the creeping bentgrass study (55 DAT) between 1300 and 1400 h during mostly sunny conditions (illuminance 50 000 lux). Images were collected by a researcher standing immediately next to the plot while holding the camera directly over the center of the plot 1.5 m above the turf canopy. Care was taken to avoid casting shadows on the turf inside plot. Concurrent to the collection of digital images, the zoysiagrass and creeping bentgrass studies were visually rated for color by five and three independent researchers (rater experience ranged from a minimum of 2 yr to >10 yr), respectively. Color ratings were based on a 1 to 9 scale where 1 = tan or brown turf, 6 = minimum acceptable color, and 9 = optimal dark green color. A DGCI was created from the HSB values to obtain a single value from digital image color analysis for comparison with values from subjective visual ratings. The index was created to measure the relative dark green color of an image using the following equation:
 |
The color index was calculated from the average of transformed HSB parameters. Each transformed parameter measures dark green color on a scale of zero to one. Since the hue of most turfgrass images ranges between 60° (yellow) and 120° (green), the maximum dark green hue was assigned as 120°. Therefore, the dark green hue transform was calculated as (hue - 60)/60, so that hues of 60° and 120° would yield dark green hue transforms of zero and one, respectively. Since lower saturation and brightness values corresponded to darker green colors, (1 - saturation) and (1 - brightness) were used to calculate the dark green saturation and brightness transforms, respectively. The average of the transformed HSB values yielded a single measure of dark green color, the DGCI value, which ranged from zero to one with higher values corresponding to darker green color.
Analyses of variance were performed using PROC GLM in SAS Statistical Software (SAS Institute, 1996) on the visual rating, HSB, and DGCI data sets. For a given color parameter, treatment and/or interaction effects were determined significant when the corresponding ANOVA f test had a P value 
0.05. In such cases, a Fisher's protected LSD test was performed to separate treatment means (Freund and Wilson, 1993).
Three digital images were taken on plots from the zoysiagrass and creeping bentgrass studies to compare the variance of digital image analysis with subjective visual rater variance. Since the visual rating scale was unrelated to color values obtained from digital image analysis, the relative variances (coefficients of variation) were used for statistical comparison. Sample variances were calculated as the within-plot mean square for each color quantification method. Confidence bounds (95%) were constructed for the sample means and the within-plot variances and were used to calculate confidence bounds for the coefficients of variation. The relative variances of the methods were determined to be significantly different if the respective confidence bounds for the coefficients of variation did not overlap.
Cultivar Color Differences
Plots from a bermudagrass cultivar trial were used to assess the ability of digital image analysis to quantify visual color differences among cultivars. The trial was established in the summer of 1997 at the University of Arkansas Research and Extension Center in Fayetteville, AR (silt loam, Typic Hapludults, pH 6.2), and was a test site for the 1997 National Turfgrass Evaluation Program (NTEP) bermudagrass trial (NTEP, 1999). Individual plots were 1.4 m2 and maintained at a 1.9-cm mowing height. The study was replicated three times in a completely randomized design.
Digital Images were taken as described previously on each replication of four cultivars that varied in green color (NTEP, 1999): ‘Cardinal’ (strong yellow-green), ‘Shanghai’ (dark gray-green), ‘Mini-Verde’ (strong dark yellow-green), and ‘Tifway’ (typical bermudagrass green color). The plots were photographed on 21 Sept. 2000 between 1325 and 1335 h during overcast conditions (illuminance 5000 lux).
One-way ANOVAs were performed using PROC GLM in SAS Statistical Software (SAS Institute, 1996) on the HSB and DGCI data sets, with cultivar as the treatment variable. For a given color parameter, differences were determined significant among cultivars when the ANOVA f test had a corresponding P value 0.05. In such cases, a Fisher's protected LSD test was performed to separate cultivar differences (Freund and Wilson, 1993).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
In both studies, the coefficients of variation for HSB and DGCI ranged from 2 to 18 times less than that of visual ratings (Tables 3, 4). All coefficients of variation for the digital image analysis parameters were statistically smaller than the CV% for the visual ratings based on the 95% confidence intervals.
|
|
10° (and significantly) lighter in hue than the other three cultivars. This result was consistent with ‘Cardinal’ appearing a lighter shade of green to the eye than the other three cultivars. ‘Cardinal’ also ranked lowest in genetic color among 28 cultivars in the 1997 NTEP trials when results were averaged across 18 trial locations (NTEP, 1999). ‘Shanghai’, which appeared darker to the eye than the other cultivars, had a significantly lower saturation level than the other cultivars (Table 5). The dark color of this cultivar was apparently due to its grayish green color (less saturation), rather than it being a darker shade of green (higher hue).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
These results confirm that visual ratings can be used to separate treatment effects on turf color. In most cases, raters ranked the turf plots similarly although differences existed in their absolute rating values. Therefore, color ratings remain a valid evaluation tool if data are not compared across raters. However, the accuracy of digital image analysis, demonstrated in the calibration experiments, enables researchers to record reflected turfgrass color on a standardized scale rather than using arbitrary rating values. Therefore, valid comparisons of color data across researchers, locations, and years are possible with digital image analysis.
Creeping bentgrass plots had significant differences in HSB levels, whereas zoysiagrass plots were significantly different only with regard to hue. This may be due to a genetic difference in N uptake and utilization between the two species. However, in both species, significant DGCI differences existed due to N treatments. Therefore, the DGCI is a more consistent measure of dark green color across species than the individual measurements of H, S, or B. Since N fertility significantly affected the HSB levels of creeping bentgrass and zoysiagrass (Fig. 5A), color measurement using digital image analysis may be capable of assessing the N status of plant tissues. For example, zoysiagrass plots exhibiting the darkest green N responses had hue angles near 90° while the most chlorotic plots had hue angles near 70°. Other research has demonstrated that correlations exist between the N content of creeping bentgrass tissue and its color, measured by colorimeter (Landschoot and Mancino, 2000).
|
The ability to distinguish color differences among turf plots as either H, S, or B differences is a significant advantage of digital image analysis over subjective visual ratings. For example, a turf that has a darker color because of grayish genetic color may not be as aesthetically desirable as a turf that is lighter in appearance but is saturated with green color. Consequently, there exists a potential for evaluator bias, which may have occurred in the 1997 vegetative bermudagrass NTEP trails where the dark grayish variety ‘Shanghai’ ranked among the top three cultivars in genetic color in 13 of the 18 test sites, while it ranked near the middle or bottom at the other five sites (NTEP, 1999). Rather than ‘Shanghai’ exhibiting different genetic color at the various NTEP locations, this discrepancy may have been due to varying evaluator perceptions of optimal dark green color for bermudagrass.
Digital image analysis was more time consuming than visual color ratings, but far less labor intensive than traditional laboratory methods that are used to quantify turf color (amino acid and chlorophyll assays). Images were collected in the field at a rate of 2 images per minute and were analyzed with SigmaScan at a rate of 3 images per minute. Although subjective ratings require less time than digital image analysis, the color data obtained from digital image analysis are free from researcher bias and inaccuracies and include information on individual HSB parameters. Furthermore, Sigma-Scan macros have been developed for batch-analysis of an unlimited number of images (Karcher, 2001, unpublished data).
Another advantage of digital image analysis over other objective color evaluation methods is the ability to measure large areas of turf in situ. The area of turf that is possible to evaluate is limited only by the height of the camera above the canopy and the subsequent field of vision. An el-shaped monopod was designed at the University of Arkansas that enables images to be taken of turf areas in excess of 30 m2 (a remote control releases the camera shutter). This is a significant improvement over standard colorimeters that typically measure areas smaller than 20 cm2 (less area than a 35-mm slide). In addition, if a turf plot is not uniformly green due to disease, injury, or dormancy, a color threshold technique can be used within SigmaScan to quantify the color of only the green portions of an image which correspond to healthy turf (Richardson et al., 2001). Another advantage of digital image analysis is that once images are obtained, they can be stored indefinitely before analysis. For instance, images of field trials can be collected regularly during the growing season, but analyzed during the off-peak months. In contrast to visual color ratings, trained, experienced researchers are not needed to evaluate turf color using digital image analysis.
Light conditions may affect the results from these techniques, although successful results were obtained under both sunny and overcast conditions in these studies. Digital color analysis may not be as effective during dawn or dusk due to increased shadows within the turf canopy. In addition, comparisons of color among turf plots from different locations and times may only be possible if the images are collected under equal light conditions. This could be accomplished through the use of standard artificial light sources while collecting images either at night or in an enclosed system.
A digital camera capable of acquiring high quality images is becoming commonplace in turfgrass research programs. The ability to capture extensive information of turfgrass in situ makes it a viable tool to quantify turfgrass parameters commonly of interest in field experiments. In addition to color quantification, digital image analysis has been used successfully to quantify percentage turfgrass cover (Richardson et al., 2001) and may potentially be useful in quantifying turf parameters such as weed infestation, disease incidence, herbicide phytotoxicity, leaf area, and recovery from injury.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 4, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Dai, M. J. Schlossberg, and D. R. Huff Salinity Tolerance of 33 Greens-Type Poa annua Experimental Lines Crop Sci., May 1, 2008; 48(3): 1187 - 1192. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Richardson, K. W. Hignight, R. H. Walker, C. A. Rodgers, D. Rush, J. H. McCalla, and D. E. Karcher Meadow Fescue and Tetraploid Perennial Ryegrass--Two New Species for Overseeding Dormant Bermudagrass Turf Crop Sci., January 22, 2007; 47(1): 83 - 90. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Schlossberg and J. P. Schmidt Influence of Nitrogen Rate and Form on Quality of Putting Greens Cohabited by Creeping Bentgrass and Annual Bluegrass Agron. J., January 1, 2007; 99(1): 99 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. DaCosta and B. Huang Minimum Water Requirements for Creeping, Colonial, and Velvet Bentgrasses under Fairway Conditions Crop Sci., December 2, 2005; 46(1): 81 - 89. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. E. Karcher and M. D. Richardson Batch Analysis of Digital Images to Evaluate Turfgrass Characteristics Crop Sci., June 24, 2005; 45(4): 1536 - 1539. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
