Australian scientists have achieved a remarkable breakthrough in the fight against fruit fly infestation in cherries and other fruits. Lead researcher Maryam Yazdani and her team at the CSIRO have developed an advanced AI-powered imaging technology that has the potential to revolutionise the detection and management of fruit fly infestation.
Yazdani explained that the new AI programmed optical scanning is specifically designed to enhance the efficiency and effectiveness of detection, particularly in horticultural packhouses where optical scanning plays a crucial role in the quality grading process. The project’s aim is to integrate an imaging system for pest detection into existing optical grading processes in packhouses, streamlining the identification and management of fruit fly infestations.
The potential impact of this technology extends beyond improving fruit grading efficiency. Yazdani expressed optimism that the new technology could facilitate access to markets that are currently inaccessible to Australian exporters due to fruit fly risks. By providing an additional tool for border security regulators to mitigate the risk of pest transportation, this innovation could play a critical role in expanding export opportunities for Australian fruit growers.
One of the key advantages of the optical scanning technology is its potential to serve as an alternative to fumigation, which is a costly and increasingly restricted method of fruit fly management. With the AI program’s capability to detect infestation, including recently laid eggs within the fruit, existing sorting technologies within packhouses can then remove the affected fruit, making the management of fruit fly infestation more precise and targeted.
Yazdani highlighted the importance of high-quality data in developing the AI model, stating that the team generated over 40,000 images during the three-year development process to “train” the program. This meticulous approach has resulted in a detection model that achieved about 95% accuracy in identifying fruit fly damage in cherries, demonstrating the effectiveness of the technology.
While the initial focus was on identifying Queensland fruit fly in Australian cherries, the technology has been tested on various fruits and fruit fly species, including blueberries, apples, and nectarines, as well as the Mediterranean fruit fly. As the testing progresses, the team aims to expand the application of the technology to other fruit types and fruit fly species, further broadening its potential impact on the industry.
The CSIRO is currently collaborating with a commercial partner to explore integration opportunities and intends to license the model to existing packhouse grading systems. This marks a significant step towards the real-world implementation of this groundbreaking technology, with the potential to transform the way fruit fly detection is conducted in the horticultural industry.
In conclusion, the development of this innovative imaging technology represents a significant leap forward in the ongoing battle against fruit fly infestation. The potential for increased market access, reduced reliance on fumigation, and more precise detection and management of fruit fly infestation make this a game-changing advancement for the Australian fruit industry. As the technology continues to be refined and tested, its future impact on fruit exports and pest management is highly promising.