Introduction

Create a New Dataset

Plant Segmentation

Image Scale

Data Export and Analysis

Summary

Downloads

AraLeafMeasure (ALM, or Arabidopsis Leaf Measurement) is a software specially designed for the purposes of measuring the area of Arabidopsis plants, and performing analysis of sequences of plant images. This software was developed for Miss Yi Zhang and her colleagues in Prof. John Turner's Lab, in the School of Biological Sciences, University of East Anglia. This Appendix briefs the major functions of the software, and shows some examples of its applications.

A Dataset in ALM consist of a number of Data-series, each of which contains a series of individual Data-samples. A Data-sample is usually the image of a number of plants, whose area will be measured and analysed. The software creates the dataset from a folder where each sub-folder corresponds to a series and contains the sequence of images of the series (see Figure 1). It also allows the user to adjust the order of samples in the series once the dataset is created.

Figure 1 Create a new dataset from a folder

The plants need to be segmented and separated in order to be measured. Firstly, all possible plants are automatically segmented from the background by thresholding the green channel of the image, where the threshold can be either specified by the user, or automatically selected using a method that minimises the within-class variance. A connected-component filter is then applied to the resulting binary image, such that all the distinct objects are labelled.

In most cases each labelled component is corresponding to one plant, and the user simply needs to click on these components to identify the plants. However the plants may overlap as they grow, and therefore one component may contain multiple plants. As a solution, the ALM software provides a set of simple yet efficient interactive tools for the user to separate overlapping plants. These tools are based on Boolean operations of regions, such as AND, OR, NOT, and XOR etc. Figure 3 shows an example of separate overlapping plants using a NOT operation.

(a) Overlapping plants	(b) "NOT" operation	(c) Separated plants

Figure 3 Separate the overlapping plants by a NOT operation

While plant area is the most direct indicator to the plant growth rate, other metrics, such as perimeter, eccentricity and colour variation, can also be computed by ALM. To calculate the absolute metric values of the plants, a scale must be specified to the image by the user. In ALM, the user can place a scale with a certain length onto the image, and thus define the actual size of the measured objects. For example, in Figure 4, we use the height of the pot (95mm) to define the scale of the image such that the measured area can be scaled to the absolute value.

Figure 4 Segmented plants and specified image scale.

Although the ALM software does not provide advanced statistical analysis of the measurements, it is able to export the measured values to the standard CSV and XML formats. These formats can then be loaded into other software (e.g. Microsoft Excel) to perform further analysis. The Measurement interface of the ALM software in Figure 5 shows the measured plant area in a grid. These values are then stored in a CSV file, and loaded into Excel (Figure 6-a). In Figure 6-b, we show an example figure of growth rate comparison, which was generated by the combination of ALM, Excel and Adobe Illustrator.

Figure 5 The measurement interface of ALM.

This software implements most of the essential functions for the segmentation and measurement of plant images. Comparing to other commercial products of the same kind, it has several key advantages. Unlike those large-scale software products that provide complicated functionalities, ALM is specific, simple, and easy to use. It is completely free (the binary executable is available on request). It has been proven to be efficient and accurate when compared with some of the commercial software (e.g. WinDIAS by Delta-T Devices Ltd).

Nevertheless, improvements are still necessary. These may include automated corresponding of plants in a sequence of images, recognition and counting of plant leaves (e.g., through a model of the emersion of new leaves for a certain type of plants).

PLEASE NOTE:
- The ALM program is not to be used for commercial gain.
- The program and use thereof will be attributed to the author where appropriate (inluding demonstrations which rely on it's use).
- This programe is distributed WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Binary WIN32 Executable (.zip)
A short tutorial for the ALM software (.pdf)
This document in Adobe PDF Format (.pdf)
Microsoft Visual C++ 2008 Redistributable Package (x86)
NOTE: The Microsoft Visual C++ 2008 Redistributable Package (x86) is needed to run the ALM application on a computer that does not have Visual C++ 2008 installed.