Well as you know as a consumer yourself, you go into a retail store and you purchase fruit on the basis of what it looks like, take it home and you have an eating experience that’s bad and the research says that you won't go back to buy that fruit for four to six weeks, so it's not an instant decision but it’s certainly important to repeat purchase.
You know the big problem has been that to test mangoes in the past for maturity, they had to cut them. So the only tests we really had were the external visual cues, you know so whether it’s full, good shoulders, shape, bit of blush colour. But the reality is there's a big variation between varieties in that area and some varieties are really hard to tell by external cues, so the only sure method we had before was to actually take a mango and cut it in half and have a look at the flesh colour. So with one of the NIR guns we are now able to walk through the orchard testing fruit, seeing if there’s enough to go through a spot pick or whether they’re all ready and train up our workers as to the visual cues as to what a mature mango looks like. It just means we can get our start date correct so that we're picking fully mature fruit that the customer is going to be happy with.
Initially, we were prompted by growers to look at estimating the quality of fruit non-invasively, its internal quality being sugar content or dry matter content and that took us down the path of measuring in-line. So you’re on a pack line, you’re sorting on colour and weight. Now we were adding in another facility, that is estimating that the sugar content or dry matter content of that fruit. We were in the fields doing the dry matter measurements and we could see the grower practice of trying to estimate fruit yields, that is how much crop was on the tree, so that they could be organised in terms of labour requirements, packhouse requirements and that was all being done manually with a hand counter, so that lead us into a new line of work looking at machine vision in the field, so rather than just machine vision in the pack house, taking it into the field to estimate crop load across the field. The next step on from that of course, having seen the fruit, is to try and reach out to pick the fruit to automate the harvest.
It’s actually getting more advanced and so now the machinery is able to identify and count fruit in the orchard and last year it turned out to be with only a few percents wrong from the actual count of numbers of mangoes are in that entire block. Now that technology is also able to give us the size range of that fruit and so knowing how much fruit is in the block, knowing when it’s going to be mature, knowing the size of the fruit means we can schedule our workforce, we can order the right number of cartons, we can order the right numbers of inserts to go in those cartons. This is going, this could be a real game-changer for not only our farm but for the industry.
Our Non-Invasive Sensor team has led the world in the use of near-infrared spectroscopy for assessing horticultural produce.
Led by Professor Kerry Walsh, the team is focused on the development of new sensor hardware and applications of existing sensors that can assess agricultural commodities and advance productivity without damaging the product.
In partnership with international technology companies and the Australian horticultural supply chain, the team has delivered new tools to accurately assess the ripeness of mango crops prior to harvest and retail. In some cases, farm performance has increased by more than 40% through early and accurate assessment of the ripeness of fruit, bolstering crop productivity through optimised harvest timing and improved fruit quality.
After enjoying widespread success through the use of NIRS to improve mango production, the team is now investigating the use of machine vision for assessing mango flowering and fruiting, and robotic harvesting techniques to overcome labour shortages and occupational risks to workers.
Specialist research skills in:
- Multi-scale Monitoring tools for managing Australian tree crops
- Sensors for measuring fruit quality standards
- Sensor refinement and testing
- Measuring dry matter and Brix in ripening mango
- Technologies in fruit sorting
- Automated technologies for fruit and flower counting
- Automated technologies for fruit picking.
Robotic auto-harvester for mangos
Our researchers developed the world’s first mango auto-harvester, with its latest prototype displaying improved fruit handling in trials at Central Queensland orchards.
The auto-harvester has turned heads within the mango industry for some time, but recently underwent substantial refinement and is becoming a more viable for commercialisation.
Read more here
NIRS for fruit quality
Our non-invasives sensors team has led the world in adapting and implementing near-infrared spectroscopy (NIRS) measurement systems to assess the eating quality of mangos and predict the ideal harvest time.
NIRS sensors and the Fruitmaps app have been widely adopted within the mango industry, laying the foundation for CQU to research in-field machine vision systems to count fruit and estimate fruit size, for fruit load estimates before harvest, allowing farmers to better plan their harvest (e.g. employing the right number of pickers at the right time).
Read more here
Kerry Walsh, Professor - Sensor Systems
The theme of Kerry Walsh’s career has been the application of non-invasive instrumentation to issues related to plant performance, and in particular, in photosynthate transport - assimilate partitioning. He has led multidisciplinary work resulting in the association of a phytoplasma with the papaya dieback disorder, and the use of near infra-spectroscopy (NIRS) for fruit quality assessment. Kerry has a practical, hands-on, capability, yet an academic perspective on life. He strongly believes that his R&D effort should result in a gain to society and that he should provide practically relevant training to undergraduates and postgraduates.
Dr Anand Koirala, Postdoctoral Researcher
Anand Koirala completed his PhD in precision agriculture at CQUniversity, working on machine vision technologies for fruit quality and yield estimation. He is interested in applying artificial intelligence and machine learning in different research domains through optimization and re-designing of existing frameworks. He has a coding ability in Python/Java/C++ for AI and deep learning object detection and image classification tasks using different algorithms (YOLO, SSD, Faster-RCNN) on different platforms (Keras, Dl4j, OpenCV).
Nicholas Anderson - Postdoctoral Researcher
Nicholas Anderson first began working on mango orchards while on a working holiday visa and was soon hooked on the fruit and studying crop forecasting at CQUniversity. He brought Mastt with him to Australia, IT experience from working with Alberta Health Services in Canada, where he received his undergraduate degree from DeVry University. Nick recently completed a PhD in precision horticulture with CQUniversity following his Masters of Research. His experience includes near-infrared spectroscopy and practical horticultural knowledge of mango production which ensures a strong focus on which technologies will deliver the best value to growers.