Back to Opportunities
Automated solutions for open-field plot grid creation (Hi-Res UAV imagery)
AgricultureWhat questions do you have about this RFP?
Get them answered by the team at Bayer Crop Science.
Background
In the realm of agricultural research, the creation of open-field plot grids is essential for structuring experiments, such as testing different crop varieties or growing conditions systematically, particularly in field trials for vegetables.
This conceptual map or grid is an important component of the Unmanned Aerial Vehicle (UAV) image processing workflow, as it is used as a digital overlay to analyze and manage data collected from those plots.
Traditionally, the creation of open-field plot grids for UAV image processing in vegetable trials has heavily relied on manual GPS measurements. However, this method is beset with challenges, notably being time-consuming, labor-intensive, and susceptible to errors. These difficulties are particularly pronounced in scenarios where high-precision transplanters are not readily available.
In these cases, seedlings are often planted manually, which can lead to significant variability in plant spacing and depth. When planting patterns are inconsistent, leading to unevenly spaced or misaligned crops, it becomes more difficult to accurately identify and delineate individual plots in the images captured by UAVs.
Without precise plot GPS measurements, the UAV image processing workflow cannot generate accurate crop traits or correctly assign those traits to individual plots.
Thus, there exists a pressing need for a more efficient, accurate, and reliable solution to facilitate the creation of open-field plot grids in vegetable trials.
What we're looking for
We are looking for an innovative and efficient automated workflow or technology that can rapidly generate a geospatially accurate grid for the entire field with minimal manual labor. The proposed solution should eliminate the need for manual GPS measurements and ensure the creation of a precise and complete as-planted grid, integrating seamlessly with the existing UAV image processing workflow.
While the options below represent potential solutions, we are open to other technologies or systems that can automatically recognize and delineate individual plots directly from UAV images or through integration with UAV data, creating a precise, geospatially accurate grid.
Solutions of interest include:
- AI and machine learning algorithms for image recognition and processing
- Computer vision and Geographic Information System (GIS) integration
- Automated GPS-equipped UAVs with RTK technology to improve accuracy of collected GPS signals
- Deep learning-based geospatial analysis of UAV images
- Photogrammetry and 3D Modeling of the field from UAV images
Our must-have requirements are:
- Automated Plot Grid Creation: Implement a system or workflow that automates the creation of open-field plot grids with minimal manual work
- Integration with UAV Image Processing Workflow: Ensure seamless integration with the UAV imaging processing workflow to accurately assign crop traits to individual plots.
- Accuracy and Precision: Prioritize accuracy and precision in plot GPS measurements to guarantee the integrity of UAV image data processing.
Our nice-to-have's are:
- Solutions that could address all the must-have requirements through a unified framework or system
Acceptable technology readiness levels (TRL):
Levels 1-9
What we can offer you
Eligible partnership models:
Sponsored research
Co-development
Supply/purchase
Licensing
Benefits:
Sponsored Research
Sponsored research: Up to $100,000 for a proof-of-concept, with additional potential funding for further development. Postdoctoral funding over 1 to 2 years within salary ranges consistent with the current market for further evaluation of potential leads, along with possible license of researcher's patented materials.Expertise
Selected partners will be assigned a company mentor to champion the project, identify gaps in expertise, and meet bi-monthly to ensure project success and timely "Go/No-Go" decision points. Partners will have access to Veg R&D scientists and data analysis experts to provide feedback and guidance as needed.Tools and Technologies
Tools and Technology: UAV equipment, piloting, and post-plot grid analysis can be conducted with Bayer Crop Science (BCS)-Vegetables R&D team.Data
Partners can access multiple datasets of manual GPS measurements for comparative analysis.Facilities and Services
Partners will work collaboratively with BCS-Veg R&D to evaluate technologies within BCS-owned and operated field plots.Who we are
Bayer’s vision of #HealthForAll, #HungerForNone drives our need to strengthen innovation capabilities in all areas of agriculture. We know we can’t accomplish this alone, so we're always interested to hear about novel, early-stage scientific innovations that can contribute to feeding the world without starving the planet. You have our commitment to take a look, match with our R&D priorities and provide you timely feedback.
Learn moreReviewers
PT
Phil Taylor
Director of Open Innovation & Outreach
DR
Dan Ruzicka
Innovation Sourcing Lead - Biotech Breeding
Q&A with Bayer Crop Science
The Q&A is now closed.
Sort by:
Most upvoted
Q.
May I ask what was the altitude the UAVs were flown at? and what was the resolution of the captured images? sub-centimetre accuracies could be challenging if the altitude is quite high and the resolution is limited.
3
A.
The flight altitude is between 15m-20m and resolution is between 5mm-1cm. Thanks for the question.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
May 21, 2024
Is this response helpful?
0
0
Q.
If granted, how could the equipment be given and used in other countries?
LZ
Lu Zhang, Research Development, Xi'an Jiaotong-Liverpool University
May 9, 2024
2
A.
Unfortunately not, sorry.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
May 16, 2024
Is this response helpful?
0
0
Q.
Will other formats of the stitched data be available along with Orthomosaic, such as point cloud and DSM?
2
A.
No, orthomosaic only.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 27, 2024
Is this response helpful?
0
0
Q.
Are there other inputs besides an orthomosaic such as a plot diagram or natural features or fiducials that could be used for template matching? Example images or data would be very valuable to provide context.
LW
April 12, 2024
1
A.
Thank you for the question. We will have the field/trial design metadata, such as how many columns and rows within this field, and plot spacing distance, number of plants, etc.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
April 26, 2024
Is this response helpful?
0
0
Q.
Is the UAV/Drone imaginary and video recordings already in place? Or which building blocks are needed: researcher (covered by Crop Science?), and the Partner providing technology for the computer vision technology?
MK
April 17, 2024
1
A.
Yes, we will have high-res UAV orthomosaic of the entire field created for this process.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
April 26, 2024
Is this response helpful?
1
0
Q.
Is there a need to grid qualify coordinates with legal description files of properties to confirm asymetrical plots and other artefacts that reduce precision?
SC
April 25, 2024
1
A.
No. I don't think we have legal description files of properties at this moment since this is just an early stage POC. Everything is internal not grower yet.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
April 26, 2024
Is this response helpful?
1
0
Q.
Could you provide more details about the typical size and layout of the fields? What software or tools are currently used for processing UAV imagery and analyzing data in your workflow?
1
A.
The typical size of the field is from 20 - 80 acres, and the shape is generally rectangle and the numbers of rows and columns can vary.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
May 9, 2024
Is this response helpful?
0
0
Q.
Is there any interest from Bayer regarding deep learning for weed detection and remediation?
1
A.
No, not with respect to this project. Thank you for asking.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
May 9, 2024
Is this response helpful?
0
0
Q.
What is the accuracy requirement for the geo-located plots? Thanks!
1
A.
Sub-centimeter, thank you!
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
May 9, 2024
Is this response helpful?
1
0
Q.
What are the vegetable crops considered for the grid creation?
HB
May 29, 2024
1
A.
Currently the crop may include broccoli, tomato, onion.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
Will field boundary (with geo-coordinates) be available for analysis?
HB
May 29, 2024
1
A.
Yes, we have ground control points to define the four corners of the field.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
Will high resolution UAV have the geo-tagging for each image?
HB
May 29, 2024
1
A.
Yes, but the input of the process is going to be stitched orthomosaic not individual UAV images.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
What will be the type of camera used i.e., RGB or Multi-spectral and will it be snapshot or line scanner?
HB
May 29, 2024
1
A.
Multispectral camera will be used.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
Will Bayer provide access to public cloud for data analysis?
HB
May 29, 2024
1
A.
We will provide data by using Bayer-proved data transfer service.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
By spatially accurate grid do you mean the point location of each plant in a field trial?
HB
May 29, 2024
1
A.
No, we don't need plant locations but the rectangular grid for each plot, as we showed in the webinar presentation.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
What are the intended use of created grid e.g., data analysis for field trial, crop phenotyping, etc.?
HB
May 29, 2024
1
A.
We're not ready to talk about that at this time. Thank you for your understanding.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 5, 2024
Is this response helpful?
0
0
Q.
Are the sub-centimenter GPS measurements of the each plant using the UAV image realized in a single moment or in a temporal imagery context?
1
A.
UAV image data is collected from a single flight mission.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 27, 2024
Is this response helpful?
0
0
Q.
If awarded, what's the expected start date?
1
A.
Felix,
Thanks for your question. This project is a present, on-going need. If awarded, we would do everything on our side to get the work going as soon as possible. I am unable to give you a specific date or month now, but we would want so start relatively quickly.
MR
Matthew Mc Reynolds, Open Innovation-Partnership Manager, Bayer Crop Science
April 9, 2024
Is this response helpful?
1
0
Q.
What does accuracy for the generated plot grid mean? With respect to what exactly? My understanding is that there is no unique ground truth grid and any grid that correctly delineates the plots can be considered correct.
1
A.
Success of any grid creation approach will be measured against a ground truth measurement collected by a mobile RTK GNSS receiver.
PT
Phil Taylor, Director of Open Innovation & Outreach, Bayer Crop Science
June 27, 2024
Is this response helpful?
0
0
Q.
Would autonomous solar-powered rovers work for mapping the fields?
KL
April 9, 2024
1
A.
Kenny, as always "the devil is in the details." Conceptually, it sounds very interesting. Once we see the proposal, it will allow us to understand the viability of such an approach.
MR
Matthew Mc Reynolds, Open Innovation-Partnership Manager, Bayer Crop Science
April 9, 2024
Is this response helpful?
0
0
Q.
Is the input geo-recitfied, stitched drone image, or should georecification and stitching be done by the solution? Do you need the output format to be a vector geojson file or the cropped raster/geotif, or both?
1
A.
Ce, Thanks for your question! The input is pre-geo-rectified stitched drone image. We do not need the solution to stitch for us. Also, the grid should be geo-referenced vector data, e.g. GeoJSON.
MR
Matthew Mc Reynolds, Open Innovation-Partnership Manager, Bayer Crop Science
April 10, 2024
Is this response helpful?
3
0
Share Opportunity
Complete
Seeking partners focused on
Aerospace
Agricultural Engineering
Agricultural Science
Agronomy
Artificial Intelligence
Bioinformatics
Computational Biology
Crop Cultivation and Harvesting
Earth Science
Geology
14 more
Machine Learning
Metrology
Optics & Lasers
Photonics
Plant Physiology
Statistics & Probability
Aerospace Engineering
Agriculture
Computer Science
Data Science
Environmental Science
Geography
Mathematics
Mechanical Engineering
Discover tomorrow's
breakthroughs today.
breakthroughs today.
Get new partnering opportunities
delivered to your inbox.
delivered to your inbox.
© 2023 Halo Cures, Inc.