Vesca

Lead Research Organisation: National Inst of Agricultural Botany
Department Name: Centre for Research

Abstract

Strawberry harvesting is a labour intensive task that depends critically on the availability of a large amount of low-cost labour. Growers are increasingly vulnerable to labour market price fluctuations and burdened by high employment overheads. Building on Dogtooth's proof of concept strawberry picking robot (developed during Innovate UK project Ananassa), project Vesca will deliver commercially viable picking performance using cutting edge machine learning and computer vision techniques to facilitate more efficient localization of target fruit (by more nearly optimal control of robot motion) and more accurate determination of suitability for picking. The project will also provide ancillary benefits such as yield mapping and prediction that are of significant importance to growers.

Planned Impact

Economic benefits for those involved at the project:
1) The results of this work will add to the capability of Dogtooth's current robotic platform to drive progress against
commercially important goals through increasing yield, quality of the picked product and picking performance. The timeline
for exploitation of this project is, July 2017, first production units, offered as a picking service to 1 customer (turnover 2017,
£25 K); 2018, V2.0 - incorporating technology developed in this project, roll out picking service to 2-3 customers (turnover
2018, £800K); 2019, make first machinery sales, capture 7% of UK market and release of V3.0 with increased picking
efficiency, first revenue from Australia (turnover 2019, £5.8 M). 2020, migrate from picking service to machinery sales
dominated income to achieve scale, roll out on limited basis in EU (turnover 2020, £15.8 M); 2021 exploration of markets in
North America (turnover 2021, £75 M.)

2) NIAB will decrease the phenotyping costs on research projects and at the strawberry breeding program and will help to
improve selection efficiency by providing superior pre-breeding material to meet grower and consumer demands for elite
lines that can be used in future breeding programmes. Phenotyping for fruit quality characteristics and shelf life potential is
time-consuming, labour intensive and costly.

3) Hugh Lowe Farms will be the first farm implementing Dogtooth's robotic platform improved the consistency of supply of
high-quality fresh fruit with an assured shelf-life leading to reduced wastage in store, greater resource use efficiency
leading to cost savings and will improve grower competitiveness nationally and internationally. Human pickers give
inconsistent results with direct consequences for profitability (punnets containing strawberries with inconsistent size or
shape or showing signs of mishandling would typically be rejected by customer QC procedures). Farmers use a variety of
training and monitoring procedur increase consistency but these greatly increase costs.
Wider impact:
1) Strawberry producers will benefit from reduced risk of failing to recruit enough pickers, a better price at market (for a
more consistent product), and reduced labour overheads.
2) The energy needs of the electrically powered robots will be mostly met from renewable sources, however, emissions
arising from flights taken by migrant workers will be very significantly reduced.
3) Breeders, biotechnologists all over the world will benefit from lower phenotyping costs as well as improved phenotyping
precision and will help improve selection efficiency by providing superior pre-breeding material to meet grower and
consumer demands.
4) Higher paid and higher skilled jobs will be created, by creating jobs in science, software development, mechanics, R&D,
manufacture, operations and sales.
5) More accurate and timely yield estimation with less uncertainty about the availability of labour will help maximising profit
while reducing the risk of unfulfilled contracts. All of this will contribute to a more reliable and lower cost supply for retailers
and consumers.

Publications

10 25 50
 
Description Surface spectral data were captured from 397 Malling™ Centenary and 450 Driscoll's® Amesti™ fruit using the bench top camera systems at NIAB EMR (600-1700 nm) and the portable spectroradiometer (350-2500 nm). Fruit quality parameters °BRIX, firmness, colour and fresh weight) were quantified and data were used to generate and compare correlation co-efficients for each quality attribute.
Fruit samples were also frozen in liquid nitrogen and stored at -80 °C until required for analysis of chemical quality attributes. These analyses will be carried out in Q3-Q4 and the data will be correlated with fruit spectral signatures acquired in Q2.
Malling™ Centenary data was analysed using two methods: 1) "leave-one-out cross validation" and 2) "holdout validation". Both methods showed similar results. Actual vs predicated °BRIX values were strongly correlated when the spectroradiometer and the NIR hyperspectral camera (r2 > 0.8) were used. Similar correlations with °BRIX were found between the average spectra from both sides and only one side in NIR region (r2 > 0.8)., suggesting that the spectral signature taken from one half of the fruit could be used to accurately predict the values in the other half of the fruit. The presence or absence of achenes did not influence the °BRIX correlation achieved using the NIR hyperspectral camera.
The correlation of actual vs predicted values of firmness was low with each system (r2 = 0.4 - 0.5), as was the correlation with Fresh weight (.r2 = 0.6). The correlations between actual and predicted values of fruit quality attributes achieved using the VNIR hyperspectral camera were low.
Results revealed that the key wavelength may be different from the spectroradiometer data due to the number of latent variables used for modelling. This will be further investigated in Q3-Q4.
Exploitation Route Our work on developing models to predict fruit quality attributes non-destructively will benefit the research community, the wider horticulture industry, retailers and the general public. The work will lead to an increased capability to consistently produce high yields of high quality strawberries with an assured shelf-life, whilst lowering production costs and reducing pre- and post-harvest waste. The results of this work will add to the capability of Dogtooth's current robotic platform to drive progress against commercially important goals through increasing yield, quality of the picked product and picking performance.
Sectors Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Electronics,Environment

 
Description The project outputs have been used by Dogtooth Technologies Ltd to underpin further development work in to soft fruit picking robots. Although a follow-on funding proposal was submitted to the IUK Emerging and Enabling Technologies call in November 2017, the proposal received a score of 74% and was not successful.. During 2018, requests by NIAB EMR for the consortium to revise and resubmit the bid were not actioned by the Lead Partner.
First Year Of Impact 2018
Sector Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Electronics,Manufacturing, including Industrial Biotechology
Impact Types Societal,Economic

 
Description Vesca 
Organisation Dogtooth Technologies Ltd UK
PI Contribution NIAB is an independent provider of top-class R&D for perennial crops, and is renowned for its effective translational research and dissemination of knowledge. Dr Else's team is responsible for measuring and correlating the surface spectral scans with fruit quality traits, defining the protocol for obtaining useful spectral signatures and developing predictive models for fruit quality attributes. NIAB are providing the scientific knowledge, facilities, and equipment needed to develop objective measures of fruit quality and "pick readiness", through both destructive and non-destructive methods.
Collaborator Contribution Dogtooth lead the project in both management and technological delivery. Vision system synchronisation (WP1), motion control and task planning (WP2 and WP4), and system integration and testing (WP5) will be tackled exclusively by Dogtooth. Whilst they will be collaborating on the fruit quality classification (WP3) with NIAB EMR and on testing and field trials with Hugh Lowe Farms (WP6). HLF will be responsible for providing the development site at their farm in the UK, and the trial site in Tazmania at subcontractors Burlington Berries, in which HFL are a partner. Marion and her team will develop fruit quality metrics and benchmark performance against human-labour
Impact The project is still active and so there are no outcomes to report at this time.
Start Year 2017
 
Description Vesca 
Organisation Hugh Lowe Farms Ltd
PI Contribution NIAB is an independent provider of top-class R&D for perennial crops, and is renowned for its effective translational research and dissemination of knowledge. Dr Else's team is responsible for measuring and correlating the surface spectral scans with fruit quality traits, defining the protocol for obtaining useful spectral signatures and developing predictive models for fruit quality attributes. NIAB are providing the scientific knowledge, facilities, and equipment needed to develop objective measures of fruit quality and "pick readiness", through both destructive and non-destructive methods.
Collaborator Contribution Dogtooth lead the project in both management and technological delivery. Vision system synchronisation (WP1), motion control and task planning (WP2 and WP4), and system integration and testing (WP5) will be tackled exclusively by Dogtooth. Whilst they will be collaborating on the fruit quality classification (WP3) with NIAB EMR and on testing and field trials with Hugh Lowe Farms (WP6). HLF will be responsible for providing the development site at their farm in the UK, and the trial site in Tazmania at subcontractors Burlington Berries, in which HFL are a partner. Marion and her team will develop fruit quality metrics and benchmark performance against human-labour
Impact The project is still active and so there are no outcomes to report at this time.
Start Year 2017
 
Description Fruit Focus 2017 and 2018 
Form Of Engagement Activity Participation in an open day or visit at my research institution
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact NIAB EMR's research work in precision growing of soft fruit crops was showcased in the Water Efficient Technologies (WET) Centre during Fruit Focus 2017. The WET Centre was officially opened by the Rt Hon. George Eustice, Minister of State at DEFRA. Four formal tours were held throughout the day, and three interviews were given by NIAB EMR staff for local television.
Project outputs from IUK 101623, 102144 and 102640 were presented at the fruit Focus Forum in July 2018, and demonstrated in the WET Centre during x4 sessions during Fruit Focus 2018.
Year(s) Of Engagement Activity 2017,2018
 
Description Industry visits to the WET Centre at NIAB EMR 2017 and 2018 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Industry/Business
Results and Impact The Water Efficient Technologies (WET) Centre
Following the completion of IUK 101623, NIAB EMR secured funding from industry partners to establish and run the WET Centre at NIAB EM. Our original business partners include Berry Gardens Growers Ltd, Cocogreen Ltd, Delta-T Devices Ltd, Netafim UK Ltd, New Leaf Irrigation Ltd. Start-up funding was also provided by Kent County Council and Southeast Water. Additional partners joining in 2018/2019 include H.L. Hutchinsons Ltd, the AHDB and Weatherquest Ltd.

The primary aim of the WET Centre is to create and maintain a UK Centre of Excellence, to support the commercialisation and sales of an integrated portfolio of precision irrigation and other leading-edge technologies for the horticultural sector. To achieve this, the Centre is demonstrating to horticultural growers how adoption of these "Best Practice" technologies can help them to optimise their irrigation productivity and financial returns. It also provides them with the necessary support and training required for successful uptake and operation. A key target is to demonstrate on a commercial scale that by combining Precision Irrigation with other leading technologies in a key crop such as protected substrate-grown strawberry, growers can achieve high yields of flavoursome phytonutritious Class 1 fruit whilst using resources more efficiently.

Visitors to the WET Centre in 2017 included the Berry Gardens Grower Research Awards Panel, Kent County Council, members of the SAI platform, and the Ferdonana project team. NIAB EMR's research on linking scientific knowledge of plant and crop physiology with innovative technologies to improve the precision, resource use efficiency and productivity of UK soft fruit production was presented and discussed.
Visitors to the WET Centre in 2018 included x25 visitors for the LEAF Innovation Centre Launch, the CEO and the KE Technical Manager of the AHDB, the AHDB Comms team, Kent County Council, HL Hutchinson Ltd, Duard Cloete of In2Food, the BerryDSS Project Consortium, ~120 visitors in x4 Demonstrations at Fruit Focus, a group of visiting Swedish Agronomists, and international clients of Netafim UK Ltd and CocoGreen Ltd.
Year(s) Of Engagement Activity 2017,2018
 
Description Vesca Quarterly Project Review meetings 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Industry/Business
Results and Impact Following the Kick-off meeting held on 12 May 2017, IUK Quarterly Project Review meetings were held for Q1 on 11 July 2017, Q2 on 9 October 2017, and Q3 on 16 January 2018. Results to date, progress against milestones and deliverables, and plans for the following Quarter were presented and discussed. The Q2 meeting was held at Hugh Lowe Farms and included a demonstration of the picking robot.
Year(s) Of Engagement Activity 2017,2018