• Research

CropQuant: next-generation crop monitoring for precision agriculture

Providing an affordable solution to automated crop phenomics, preventing crop losses and contributing to food security.

Project summary.

Start date: 1 April 2016

End date: 31 March 2017

Duration: 18 Months

We are working to produce an infield phenomics solution to automate and improve phenomic data capture during plant growth. Designed to be reliable and affordable, our solution can help relieve the phenotyping bottleneck and provide insight into key traits affecting crop yield within a working field environment. The aim of this system is to inform and advise breeders and scientists about the effect of environmental factors on the growth of crops, helping to provide greater yields and contributing to better food security.

CropQuant can capture high resolution growth data on monitored crops during the length of the season, allowing us to compare the changes and progression of plants to captured environmental data available through in-built sensors. This system of regular data capture allows sophisticated modelling techniques to identify trends and correlations throughout the various growth stages. Manual crop monitoring is time-consuming, expensive and greatly subject to human error, the responsibilities of which can be reduced significantly by the automated devices we are developing.

Details.

In order to produce affordable, resilient infield workstations, CropQuant has been build on top of the Raspberry Pi single board computer. The Pi is a small, low-cost and low-power board which runs a version of the linux operating system, allowing high level software and algorithm design to be implemented. Software on the Pi is written in the Python programming language, and interfaces with a powerful RGB camera module which is used to take regular, high-definition photographs of growing crops for analysis.

Using the Earlham Institute’s High Performance Cluster environment, RGB images captured by the device’s camera are formatted and processed using Computer Vision algorithms. Crop height, greenness and growth rate are monitored as well as environmental conditions and complex, non-linear models are generated from the collected data. These models can be used predictively to inform breeding decision and experimentation to enhance yield.

Part of the main design is the inclusion of digital sensors which are incorporated into the CropQuant’s enclosure. These sensors are connected to the Raspberry Pi via inbuilt digital GPIO pins and the conditions they monitor are sampled at regular intervals. Data collected from the infield workstations is collated by a central server which acts as a monitoring system for the network of devices. Management and installation monitoring is performed through the central server, which provides summary data and status information collected from throughout the season.

Tools.

Image analysis and computer vision has been performed on the Earlham Institute HPC, allowing fast and efficient processing of data at the end of the growing season.

CropQuant is an internally developed technology based around the Raspberry Pi, a commercially available single board computer and additional digital sensor modules.

Impact statement.

CropQuant aims to provide a robust and affordable crop monitoring platform for use within a working field environment. It can be used by crop breeders to evaluate crop lines and track their development through environmental changes and can provide a vital tool within the academic phenotyping pipeline.

Media coverage

Eastern Daily Press

Cambridge Network

New Anglia

Insider Media

Farming UK

EurekAlert

Phys.org