North by Northwest Partners Background
In 2020, the global banana industry was estimated to be worth $25 billion. A total of 15% of bananas produced are traded in the international market. Bananas as a crop can be tricky to cut into clusters (a collection of banana fingers) for consumers without creating substantial waste. This is because a single plant produces a bunch, which in itself can contain up to 20 ‘hands’ of bananas – these are groups of individual bananas from which clusters are cut and then packed into boxes to be shipped.
Clusters are a collection of banana fingers and the specifications of each type of cluster are unique for every customer. Orders vary on minimum and maximum width, length and number of banana fingers. Growers have to cut and pack the fruit to fulfill the specific cluster requirements and number of boxes for multiple customers at the same time. An estimated 5%-15% of the total produced fruit is wasted at the packing stage when off-cuts are discarded if they do not fit the specific order requirements.
Technology Overview
is a cutting-edge web-based decision support system offering precision agriculture. It uses state of the art artificial intelligence and 3D image processing algorithms providing optimised cutting strategies and fruit profiles tailored to each farm.
Stage of Development
The researchers currently have a prototype of OPTIcut (TRL6-7) which has been extensively and successfully trialled and validated by the leading worldwide banana producers Chiquita, Dole and Fyffes.
EFFORT LEFT: The main concern from the producers is the collection of the data needed to run OPTIcut. This is labour intensive, prone to errors and destroys the fruit. Therefore, the fruit used to collect this data, goes to waste. A state-of-the-art 3D image-based system has been developed to automatically gather data without damaging the fruit. A field validation of OPTIcut’s novel 3D image-based system is needed. A large farm in Costa Rica has agreed to act as a pilot farm for the 3D imaging system. 3D and HD 2D images will be collected and used to build a deep learning model to help visually segment the fruit. This makes measurements more accurate and enables OPTIcut to automatically detect some of the fruit damage. The team currently has four producers lined up to undertake trials of OPTIcut in their farms.
The next steps for OPTIcut include further field validation of the novel 3D image processing system. The researchers have an agreement with a large grower in Costa Rica (6000 hectares) to further trial and validate OPTIcut. This farm will act as a pilot farm where other producers can come and see OPTIcut working. As part of an Innovate UK ICURe market validation project, other growers in India, Belize and the Philippines have also agreed to participate in trials of OPTIcut.
Benefits
Trials on banana plantations in Central America using cutting strategies optimised by OPTIcut proved successful in reducing waste by 5-10%, compared to in-field estimations by experienced croppers with more than 30 years’ experience. In an industry worth around $25 billion a year generated in exports, this waste reduction adds up to more profit for the farmers.
The huge collection of data, compiled with the help of OPTIcut, is also able to help growers prevent fruit damage and further reduce waste.
This is calculated by providing them with detailed geographical information on where, historically, the main damages are occurring in the farm and the percentage of crops affected. This data also provides growers with a data-driven decision support system towards the development of managerial strategies tailored to each individual farm.
OPTIcut further helps farmers to better estimate the amount of fruit available in every site of their farm for each class of cut that they commercialise, also known as its fruit profile. This will enable them to select the most suitable site for each type of cluster.
This has in turn improved the number of high-quality fruits that are delivered to customers, meaning farmers are paid a higher price for their produce and suffer from fewer penalties for supplying either damaged goods or goods that are not compliant to the specifications.
Traditional methods of estimating fruit profiles involve cutting every single banana ‘finger’ off the banana hands. The majority of this fruit will then become waste as there is little market for single banana fingers. Instead, OPTIcut is able to use 3D imaging to measure and count the fruit, without waste or damage, higher accuracy and a lot faster than the manual alternative.
Over five years, a number of successful field trials have been performed in Costa Rica in collaboration with three of the largest global banana growers. These trials earned extremely positive reviews from growers, who are able to access the information needed on an easy-to-use dashboard, which gives them the information they need at a glance. They also resulted in a large number of enhancements of the software to better fulfil the needs of the growers. OPTIcut is a product that can have a high socio-economic impact helping banana growers reduce waste and help the environment.
Applications
OPTIcut is aimed at the agricultural sector. After the Innovate UK ICURe journey, the following potential applications were identified:
The fruit production business
Packing facility optimisation
Harnessing data
Estimating Production
Robust fruit quality control
Non-invasive fruit damage detection and tracing
Fruit profile building
Opportunity
The team is currently looking for investment to further develop OPTIcut. It is estimated that the first customers will come after three month trials at our pilot farm in Costa Rica.
OPTIcut will also be beneficial to other stakeholders such as agrochemical and water irrigation suppliers.