Fruit maturity: developing an application to assist with the decision to pick

This article provides an overview of work conducted by Central Queensland University (CQU) and Felix Instruments teams over the 2016-17 season to provide maturity data in an easy to visualise format. As a result of this work a website has been developed, this article outlines what can be found on the website, which can be viewed at: Enter ‘test’ as the username and password in all security screens.


The decision to pick can be made on several criteria:

  • internal flesh colour
  • external colour and shape
  • heat sum from flowering
  • dry matter content
  • fruit size

Internal flesh colour requires a destructive assessment, and while external colour and shape can be useful, at least in some varieties, assessment requires a level of training and consistency that may be lacking in harvest crews. Heats sums are an established technique for mango, but its use requires some diligence in keeping farm temperature records, or use of a more distant temperature record (for example as on the Northern Territory Department of Agriculture website,

In short, a tool that brings information (relevant to making the decision to pick) together in an easy to access format should have value to a grower, especially for a larger grower managing many tree blocks.

The ‘maturity’ web application (web app) (web address as shown above) attempts to bring the various tools in maturity estimation together. The science behind indices such as heat sums and the dry matter (DM) target specification is well established. The web app simply provides access to this type of information in an easy to understand format. However, the web app does rely on the following two assumptions (below):

Features of the web app

The web app has been used by a small number of growers through the 2016-17 season, and developed in response to user comments. Your further comments are appreciated. It contains the following features:

Dry matter: These features on the web app utilise DM readings from the F-750 Produce Quality Meter, with its GPS records.

The ‘Farm DM’ tab (Figure 3) displays a map showing location of individual records, with colour change from red to blue if the measurement exceeds the user selected target value. The block colour will change from red to yellow to green when the sampled fruit values fit the ‘% of fruit above target’ criterion. Individual measurements are displayed in a pop out box by clicking on dots.

Figure 3. The 'Farm DM detail' tab

Figure 3. The 'Farm DM detail' tab

The ‘Block DM detail’ tab (Figure 4) displays data of a single block, as selected in the left hand panel. Block mean DM is presented, as well as the estimated date to achieve the criteria that the user sets in the left hand panel (for example 90% of fruit above 15% DM). A graph of mean DM by date of measurement is also displayed, as well as the calculated rate of DM increase per day.

Figure 4. The ‘Block DM detail’ tab.

Figure 4. The ‘Block DM detail’ tab.

The ‘DM Table’ tab displays a list of all blocks of the farm, ordered by DM level and recommended date of harvest (i.e. date at which user set criterion on DM level is met), from the user set rate of increase. We suggest minimum input of a survey of fruit across a block, followed by one follow up measurement to get a rate of increase.

Security: Data for each farm is quarantined, available through password access.
Add polygon: This tab allows the user to enter boundary points and names for blocks within a farm.

Upload: This tab contains settings for upload of DM data, temperature data and fruit size data.
Heat units: A temperature sensor should be located on farm in a tree no further than 100 metres from the receiving computer, with temperature logged and provided to the heat sum calculator on the web app. Alternatively, the user can access data of the closest Bureau of Meteorology website. The user must enter variety and confirm required heat units, a graph of accumulated heat units to date for the current year and for the average of a 10 year historical period is displayed. A prediction is made of the date of fruit maturation based on a projection of the current year’s temperature, relative to the historical average.

Fruit size: A mobile phone application is under development (Figure 5) which calculates fruit weight from an image taken of the fruit on the tree. The user must sample a number of fruit through a given block, taking a picture of the fruit against a board. This data and the phone GPS data is transferred to the web app for display by block. Data is displayed as a histogram of fruit weight or tray size class.

In the future, fruit size data of a greater number of fruit per tree may be taken from the machine vision estimates via a system moving through the orchard.

Figure 5. A fruit sizing mobile phoneapplication is currently in development.

Figure 5. A fruit sizing mobile phoneapplication is currently in development.

Crop load: The user can record data on the percentage of growing points that have flowered by block/variety by date. The average of this data is plotted to inform a selection of flowering event dates. A second table allows entry of fruit number (estimated after stone hardening stage) associated with each flowering event to be entered on a per block basis. Fruit numbers are summed up across blocks for each flowering event and displayed with the anticipated harvest date.

Flowering and fruit number data are based on manual estimates at present, with potential to have information inputted from machine vision estimates in the future.

We propose a mobile phone application in which the user records their manual count of fruit per tree, for a selection of trees across each block. Count data and GPS records would be transferred wirelessly from the phone when in office, uploading to the web app for calculation of fruit per block given information on number of trees per block.

Availability of the maturity web app

Your comments/suggestions are welcome, please email
The web app will be available to an expanded number of users for the 2017/18 season. Thereafter it would be maintained as a commercial service under a subscription fee.
Article submitted by Professor Kerry Walsh, MSc candidate Nicholas Anderson and postdoctoral fellow Dr Zhenglin Wang from CQU, Rockhampton, and Ryan Lerud from Felix Instruments, with input from Martina Matzner of Acacia Hills Farm.

Acknowledgements: We appreciate the continuing support of Acacia Hills Farms, Groves Grown Tropical Fruits, MMM Mangoes and Simpson Farms Pty Ltd. This work was supported by Hort Innovation project MT14048, Multiscale monitoring of tropical fruit trees.