トマトアナライザ：2次元オブジェクトからの正確で詳細な形態学的および測色データを収集するために有用なソフトウェアアプリケーション

To use tomato analyzer. Tomato fruits are prepared and cut to collect digitized images. The fruit is placed on the scanner with the cut sides down and scanned to obtain a JPEG image.

Next, the images are viewed and analyzed in the tomato analyzer software. All fruit shape measurements are calculated based on the fruit boundaries recognized by the software, whereas the color test module is designed to quantify the color parameters inside the boundaries. Finally, the data is exported to an Excel file in batch mode with more than 100 images at one time or as individual images.

Hi, I'm Estevan Dena and I'm on the faculty at the Department of Horticulture and Crop Science at Ohio State University. I'm Matthew Robbins in the Francis Lab in the Department of Horticulture and Crop Science at the Ohio State University. I'm Gu Rodriguez from the Today we will show you a procedure for collecting and analyzing the shape, size, and color data of tomato fruit using tomato analyzer.

We use this procedure to study the morphology and color of tomato fruit, So let get started. Tomato analyzer, abbreviated TA is a software program that measures 37 attributes related to two dimensional shape and can also capture color data in RGB and the L star, A star B star universal slab color space in a semi-automatic and reproducible manner. Here TA is used with scanned images of tomato fruit.

To begin this procedure, make sure the fruit is clean and dry. If the fruit is fleshy, make sure it is not over ripe because softening may cause shape deformation when cutting the fruit. In order to measure how flat or elongated a fruit is and for all attributes except those under the latitudinal section category, use a serrated knife or a new razor blade to cut the fruit longitudinally through the center to measure the thickness of the peri carp and for all attributes under the latitudinal section category, as well as some under basic measurement and morph cut the fruit transversely.

Color analysis can be applied to longitudinally or transversely cut fruit, as well as other sections designed to highlight specific features. Finally, dry the internal part of the fruit by blotting with tissue or paper towels. If the fruit is very juicy, the cut fruit is now ready for image collection.

For high throughput analysis, it is essential to scan fruit from only one plant or genotype at a time and to scan fruit that is either longitudinally or transversely cut, but not a mixture of both on the same scan. Start by placing the cut fruit on the scanner with the cut side down carefully and precisely. Place the fruits on the scanner close together but not touching.

The better the objects are aligned with minimal large empty spaces between them, the less manual adjustment with imaging software such as Adobe Photoshop is necessary. Prior to analyzing the fruit in ta, place a ruler above the fruit used to verify that the correct scan resolution has been selected, which is important for the accurate measurement of fruit attributes. Place a label below the fruit to verify the file is named correctly.

Since the TA software uses image resolution to accurately measure size, proper selection of the scan resolution is important as a rule for objects between one to eight centimeters. Scan at 300 DPI for objects larger than eight centimeters. Scan at 100 DPI and for very small objects such as seeds scan at 750 DPI or higher.

For batch analysis where the fruit size is not highly variable among different plants, select one scanner resolution setting for all images to be collected for the same experiment. Also set the output image size at the highest number of colors available in the scanner software. Create a dark background for the objects by placing a cardboard box on top of or over the scanner screen.

A black or very dark background prevents shadows that interfere with the analysis. A white background will cause the TA to crash since it is optimized for the use of a dark background. Light reflection and background may also be recognized by the software scan.

The fruit after the initial scan cropped the image before saving it to avoid large empty spaces. If the image is not cropped at this stage, it must be cropped using imaging software such as Adobe Photoshop prior to analysis with ta. Save the scanned image as a JPEG file for color analysis, which involves translation of the RGB values to L star, A star and B star parameters calibrate the scanner using a previously prepared color checker.

Here a custom color checker is used that represents the range of colors in the images of interest. A standard color checker can be used as well. Scan the color checker during the scanning of the fruits.

Save the color checker scan in the same folder as the fruit images scanner. Calibration is necessary because scanners may differ in how they capture and interpret. RGB Color calibration allows the conversion to universal color space proceed to image analysis.

Tomato analyzer version 2.200 requires a Windows operating system version 2000 or higher and can be downloaded from our website. A user manual and the color test manual are also available from the same website to make sure that the original images are saved in the same folder as the adjusted images produced by the TA application. Place the folder with the original images on the hard drive and not in the My Documents folder or on a remote server.

Launch the TA program. First set the DPI and measurement units under the settings menu by selecting scanner DPI. The DPI setting must be the same as the image files so that the size measurements are accurate.

The units used in the dialogue box determines the units for the data output. In the same dialogue box, adjust the brightness if required when the objects are relatively dark. Next, select the attributes to be measured from the settings menu by selecting measurement saved.

The attributes are grouped in 10 categories from basic measurement to morpho metrics. One can select or deselect individual attributes or an entire measurement group by clicking on the group or attribute. All attributes of a measurement group can be shown by clicking on the plus sign.

The selected attributes are displayed by category in the window in the bottom right corner of the screen. After setting the input resolution and the output units and selecting the attributes to be analyzed, click on the open image button and select the image file. From the pop-up dialogue box, the selected image will be displayed in the left window.

To analyze the open image, click the analyze button. When finished, the perimeter of each fruit is highlighted with the yellow line and the data are displayed in the lower right data window. Note that the data for color are not shown in this window.

The software automatically deselects very large or small objects such as a ruler or label. The deselected objects are highlighted with a blue line. Use the ruler to make sure that the units and resolution were selected correctly.

By comparing with the height and width measurements, additional fruits can be deselected by right clicking on their image only. Items outlined in yellow are displayed in the data window and exported items outlined in blue are not included in the analyses. Individual objects can be displayed in the upper right window by left clicking on them with the mouse.

The data in the lower right data window correspond to a particular object in the image and are displayed in the same order as the objects in the image. The first row displays the values for the object in the upper left hand corner and the last data row displays the values for the object in the lower right corner of the image. In addition to clicking on a fruit, to highlight the corresponding data row, one can click on a row to display the corresponding fruit in the top window on the right.

By clicking on the attribute tab in the lower right window, it is possible to view how the attribute is being measured for each particular fruit. For example, by clicking on distal end angle, the angles on the fruit in the left window that will be measured by the software will be shown. This feature is very useful when identifying objects that require manual adjustment.

For more accurate measurements, a click on the revised button shows the manual adjustments that can be made in the application. Following are some of the most frequently used manual adjustments. First left click an object in the left window image if it needs manual adjustment.

If the boundary needs adjustment, adjust this feature first select boundary. From the dropdown menu of the revised button, select the boundary location to be modified by left clicking on the start point and endpoint of the incorrect boundary. As a result, the delimited boundary will be removed.

To add a new boundary left click from the start point toward the end point, continue clicking to follow the desired contour. To confirm the new boundary, press the enter key. Otherwise, press escape key to cancel this operation.

Objects that require rotation will need to be adjusted after the boundary, but before the remaining manual adjustments, click on the appropriate fruit in the left window. Click the arrow next to the revised button and select rotate from the dropdown list and access will be displayed in the upper right window. Drag the green square at the end of the axis and the fruit will rotate accordingly.

Press the enter key to finish this adjustment. Other manual adjustments are available. Instructions are found in the accompanying written protocol.

For some attributes such as angles and blockiness, select the settings at which the measurement should be taken. The angles can be calculated at various positions from the ends of the fruit. These settings can be chosen under the settings menu.

Blockiness is calculated as the ratio of the width at a user selected proportion of the height closest to the distal or proximal end of the fruit to the mid width. For color test, the scanner needs to be calibrated. First, open the scanned color checker image and analyze it in the same way that was demonstrated for the fruit.

To collect L star, A star B star values, make sure TA recognizes each patch as an object to analyze marked by the yellow boundary. If using a standard color checker, the software may not be able to recognize the darker swaths. These colors will not be included in the calibration.

Swath recognition can be improved by changing the brightness under the scanner DPI from the settings menu. Then under the settings menu, select color test for calibration. Default settings should be used when the dialogue box appears.

Make sure the correction values are set to one for the slope in the left boxes and zero for the y intercept in the right boxes. Minimum blue value should be set to zero, and parameters one and two can be ignored. Now click on the analyze button within the color test dialogue box.

A new window appears for saving the output in an Excel document. Specify the name and the directory for the data file. The output data file will contain the RGB values, star ASTAR B values and calculations for HU and chroma.

For each color patch, plot the L star ASTAR B values for each color patch against the colorer values for L star, A star and B star, which are available from the manufacturer of the color checker or in the accompanying written protocol. Determine the regression equation and record the slope and y intercept for each parameter. In the color test dialogue box, enter the inverse of the slope and the reverse of the sign of the y intercept value.

For L star, A star and B star, these values will be used as correction values. For the color test settings. In this dialogue box, set the minimum blue value to 30 for tomato fruit.

If the fruits are smaller than three centimeters, set the minimum blue value to 20. This value may need to be adjusted if TA has trouble finding the correct boundaries or if analyzing objects other than tomato fruit. The color test also allows the user to define two parameters that report the percentage of pixels that fall into specified ranges.

A few values for analysis of tomato set the lower and upper values for parameter one at 70 and 100 respectively, which correspond to undesirable yellow and green. Yellow flesh color. Set the lower and upper values for parameter two at zero and 50 respectively.

These correspond to the desired red color. Save the settings for these values. These values are then applied to all images that are analyzed until the program is closed.

For scanned images, select the Illuminate C two degree option and not the Illuminate D 65 10 degree option, which is only applied for images acquired in natural light After calibration. Analyze the fruit images for color after manual adjustments of an image and image analysis by TA detailed in the accompanying written protocol. Click on the save fruit button.

All current information, including manual adjustments and deselected objects are saved in a new file with the same name as that of the original image. But with the TMT extension to revert to the original image file without any of the adjustments, simply delete the associated TMT file. Alternatively, select analyze and reanalyze the original image.

However, note that any TMT file with the same name as the scanned image remains associated with the original image unless one overwrites to save changes. Export data by selecting the export button, the data are exported to an Excel file. The attribute values for each fruit, the average and the standard deviation are displayed.

The batch analysis function allows exporting attribute values from two or more images. Start up the batch analysis by clicking on the open image button. Select the image files to be batch analyzed.

Multiple files are selected using the shift or control key while selecting additional files. After the files are selected, click on the open key one can then select the type of batch analysis output average, only. Average and standard deviation or individual measurements per image.

Choose a name for the Excel file that will be created and click on the save key. The software automatically opens the files and begins the batch analysis. If the image files have been previously analyzed and saved by ta, the saved TMT files will be opened for the batch analysis.

If the files have not been previously analyzed, the software performs the analysis of images without manual adjustments and DES selections, TA can perform a batch analysis of at least 100 images at 600 DPI. Batch analysis can also be used for the color test up to 100 fruit images can be analyzed in batch depending on the computer's hardware. Open the color test option under the settings menu.

Check the batch analysis and click analyze. A new window appears to select the images to analyze. In the next window, specify the name and the directory for the data file.

Click save and the batch analysis of the color test begins. Here are the shape and color results for two images representing variation in tomato. The output files of both images are on the right displaying the values of several shape and color attributes for each fruit.

The results of batch analysis are displayed in this output file showing the average attribute shape values and the standard deviation for 10 tomato varieties. The color analysis output of batch analysis contains average color values for each fruit, as well as other parameters. We've just shown you how to collect tomato fruit images and analyze them using the most commonly used features available in tomato analyzer.

Although this software was developed to analyze tomato fruit, it can also be used to analyze other plant organs, including leaves and seeds and tomato and other plant species. So that's it. Thanks for washing and good luck with your experiment.