At Hiphen, we accompany our clients towards each step of their plant phenotyping journey with the goal to provide a hassle-free experience and to deliver excellence in data quality. Drone data acquisition is often the starting point of your phenotyping project, and if not executed correctly, your agronomist team is likely to receive poor quality results. The Hiphen Academy is all about limiting this risk.
Like most of the agro-companies we serve around the world, you are looking at internalizing field data acquisition, which means training your own crew to fly drones across your field trials. We believe that it is the best way for you to reach a viable scale vs. economic balance, but in this process clients have to consider the following 5 criteria:
Well, the Hiphen Academy is your very own e-learning platform that has been developed to provide answers to these questions. It contains all the in-depth knowledge we gathered over the past 7 years of in-field missions worldwide, which took us almost 2 years to package cleverly in a user-friendly online learning platform.
When we started designing this resource, we quickly realized that it was important to emphasize that it is not aimed at teaching how to fly a drone – drone is easy, anyone can operate a drone these days – but rather at explaining how to fly a drone for agriculture!
Many times in the past we ended up exchanging with clients who had just outsourced drone acquisition to an external provider, expecting this partner to expertly deliver good data. However, some clients had bad experiences since the drone pilots did not have a sound understanding or sensitivity about the parameters and protocols they should be using to measure plant features. Drone flights should not all be treated as equals if you are interested in computing wheat head density than if you want to assess plant height or vegetation indices – which are traits that can be achieved with less-demanding KPIs.
The Hiphen Academy contains 6 courses and 45 lessons dedicated to drone plant phenotyping, representing 8+ hours of training, all based on real-life experience. You can consume this premium content at your own pace to kick things off, but the story doesn't stop here. It is invaluable to also provide you with the ability to ask questions to our team of experts and to receive their feedback on your very own data acquisitions to guide you all the way through until you become fully autonomous.
The Hiphen Academy is the first and the only one e-learning platform of its kind dedicated to crop researchers from plant breeding or crop protection companies around the world, who want to get the most of their field experiments data. The following video will introduce you to this intuitive and user-friendly e-learning platform:
As Alexandra explained in this video, the Hiphen Academy is composed of 6 courses that cover in detail all the know-how required before, during and after the drone flight. This knowledge is accessible via a single e-learning platform built to let you learn flawlessly and at your own pace.
A full demo of the Hiphen Academy has been made in a previous webinar. You can watch the replay HERE
Feel free to get in touch should you have any questions via academy@hiphen-plant.com
Your Hiphen Team.
Based on a quick survey conducted on social medias recently, the vast majority of agtech professionals (72%) think that drone is today the most versatile and easy to use device for plant phenotyping. It is true that, the latest improvements in drone technology makes this type of systems more performant and more accessible than ever. Thus, most of the crop researchers we work with around the globe are internalizing drone data acquisitions to make high-throughput plant phenotyping part of their plant assessment routine.
With modern technology, flying a drone is easy. Almost anyone can now get their hands on affordable equipment that you can familiarize yourself with fairly quickly to take-off and land by the press of a few buttons on a tablet. Once data is acquired, Hiphen then provides a hassle-free experience in the sense that all you have to do is upload your data on our cloud platform and then download your results – all the heavy lifting machinery in between is our responsibility. What one gets in return from drone acquisitions is an extensive list of valuable agronomic traits that accurately and objectively describe and assess plant architecture, behavior and sanitary state. You can consult our drone phenotyping catalog for more information about the traits that can be computed.
Equipment from manufacturers such as DJI are reliable for plant phenotyping, and ensure a frictionless experience with more and more automation and fewer interactions between the pilot and the machine. That being said, while most agtech experts agree that flying a drone is easy, flying a drone for agriculture requires special knowledge about the appropriate sensor choices and configuration, flight parameters and protocol to ensure that you acquire top-quality data.
The most common mistake is to acquire images with insufficient resolution. The compromise between speed and resolution (i.e. the number of pixels per ground cm) is critical and choosing inappropriate sensors and altitudes can result in poor data quality. Depending on the traits you are interested in analyzing, these parameters have to be carefully specified in advance.
Front and side overlap ratios ensure that each of your trial plots will be photographed multiple times to be able to generate an accurate field map – often referred to as an orthomosaic in our photogrammetry jargon. Setting the wrong overlapping parameters is very likely to result in data gaps in parts of your field trials.
GCPs are targets placed on the ground that have to be geo-referenced and hooked so they remain in the same precise spot during all the flights of the campaign. Not using this common practice is likely to result in inaccurate field maps which will make time series analysis almost impossible.
You are measuring plants, and plants are living organisms. Depending on your crop, genotypes, experiment, and the traits that have to be measured, it is important to acquire top-quality data at the appropriate phenological stage. In parallel, certain weather conditions have to be carefully anticipated to avoid blurry, overexposed or underexposed images, or even multispectral image processing being negatively impacted by wet field conditions.
As mentioned above, the choice of sensor is critical to deliver the appropriate image resolution, along with the orientation angle of the camera. Setting the wrong sensor parameters and not orientating your camera at NADIR is likely to negatively impact the quality of your drone data.
At Hiphen, we are present at every step of your plant phenotyping journey, starting with data acquisition, then data processing and data analytics to serve your applications such as best cultivar selection, yield prediction, ideotype qualification, trial quality assessment, cultivar risk assessment, and so on.
Thus we accompany you with your data acquisition by providing you with a drone acquisition protocol document that lists all the flight parameters and guidelines your crew should follow to ensure data acquired is high quality and consistent from one pilot to the other. In addition, we are delighted to announce the launch of the Hiphen Academy this fall. The Hiphen Academy is the first e-learning platform of its kind, fully dedicated to crop researchers focused on drone plant phenotyping projects. This online resource contains 6 courses and 45 lessons dedicated to drone plant phenotyping, representing 8+ hours of training, all based on real-life experience. Here is a quick video to introduce you to the Hiphen Academy.
👇 Watch the replay of our Webinar about the Hiphen Academy 👇
You can also register HERE to grab the latest news about the Hiphen Academy.
So stay tuned by visiting our website and LinkedIn page to receive the latest news about Hiphen and understand how we can help you to achieve your plant phenotyping goals.
Your Hiphen Team.
Several crops bear reproductive organs (RO) at the top of the canopy after the flowering stage, such as ears for wheat, tassels for maize, and heads for sunflowers. RO present specific architecture and optical properties as compared to leaves and stems, which may impact canopy reflectance. This study aims to understand and quantify the influence of RO on the bi-directional variation of canopy reflectance and NDVI.
Multispectral camera observations from a UAV were completed over wheat, maize, and sunflower just after flowering when the RO are fully developed and the leaf layer with only marginal senescence. The flights were designed to sample the BRDF with view zenith angles spanning from nadir to 60°and many compass directions. Three flights corresponding to three sun positions were completed under clear sly conditions. The camera was always pointing to two adjacent plots of few tenths of square meters: the RO were manually removed on one plot, while the other plot was kept undisturbed.
Results showed that the three visible bands (450 nm, 570 nm, 675 nm), and in a lesser way the red edge band (730 nm) were strongly correlated. We, therefore, focused on the 675 nm and 850 nm bands. The Bi-Directional Reflectance (BRF) of the canopy without RO shows that the BRF values were almost symmetrical across the principal plane, even for maize and sunflower canopies with a strong row structure. Examination of the BRF difference between the canopy with and without RO indicate that the RO impact canopy BRDF for the three crops. The magnitude of the impacts depends on crop, wavelength and observational geometry. These observations are generally consistent with realistic 3D reflectance simulations. However, some discrepancies were noticed, mainly explained by the small magnitude of the RO effect on canopy BRF, and the approximations made when simulating the RO layer and its coupling with the bottom canopy layer. We finally demonstrated that the RO layer impact the estimates of canopy traits such as GAI as derived from the multispectral observations.
Early-stage plant density is an essential trait that determines the fate of a genotype under given environmental conditions and management practices. The use of RGB images taken from UAVs may replace traditional visual counting in fields with improved throughput, accuracy and access to plant localization. However, high-resolution (HR) images are required to detect small plants present at early stages. This study explores the impact of image ground sampling distance (GSD) on the performances of maize plant detection at 3-5 leaves stage using Faster-RCNN. Data collected at HR (GSD=0.3cm) over 6 contrasted sites were used for model training. Two additional sites with images acquired both at high and low (GSD=0.6cm) resolution were used for model evaluation. Results show that Faster-RCNN achieved very good plant detection and counting (rRMSE=0.08) performances when native HR images are used both for training and validation. Similarly, good performances were observed (rRMSE=0.11) when the model is trained over synthetic low-resolution (LR) images obtained by down-sampling the native training HR images, and applied to the synthetic LR validation images. Conversely, poor performances are obtained when the model is trained on a given spatial resolution and applied to another spatial resolution. Training on a mix of HR and LR images allows to get very good performances on the native HR (rRMSE=0.06) and synthetic LR (rRMSE=0.10) images. However, very low performances are still observed over the native LR images (rRMSE=0.48), mainly due to the poor quality of the native LR images. Finally, an advanced super-resolution method based on GAN (generative adversarial network) that introduces additional textural information derived from the native HR images was applied to the native LR validation images. Results show some significant improvement (rRMSE=0.22) compared to bicubic up-sampling approach.
The Global Wheat Head Detection (GWHD) dataset was created in 2020 and has assembled 193,634 labelled wheat heads from 4,700 RGB images acquired from various acquisition platforms and 7 countries/institutions. With an associated competition hosted in Kaggle, GWHD has successfully attracted attention from both the computer vision and agricultural science communities. From this first experience in 2020, a few avenues for improvements have been identified, especially from the perspective of data size, head diversity and label reliability. To address these issues, the 2020 dataset has been reexamined, relabeled, and augmented by adding 1,722 images from 5 additional countries, allowing for 81,553 additional wheat heads to be added. We would hence like to release a new version of the Global Wheat Head Detection (GWHD) dataset in 2021, which is bigger, more diverse, and less noisy than the 2020 version. The GWHD 2021 is now publicly available at this http URL and a new data challenge has been organized on AIcrowd to make use of this updated dataset.
Plants density is a key information on crop growth. Usually done manually, this task can beneficiate from advances in image analysis technics. Automated detection of individual plants in images is a key step to estimate this density. To develop and evaluate dedicated processing technics, high resolution RGB images were acquired from UAVs during several years and experiments over maize, sugar beet and sunflower crops at early stages. A total of 16247 plants have been labelled interactively. We compared the performances of handcrafted method (HC) to those of deep-learning (DL). HC method consists in segmenting the image into green and background pixels, identifying rows, then objects corresponding to plants thanks to knowledge of the sowing pattern as prior information. DL method is based on the Faster RCNN model trained over 2/3 of the images selected to represent a good balance between plant development stage and sessions. One model is trained for each crop.
Results show that DL generally outperforms HC, particularly for maize and sunflower crops. The quality of images appears mandatory for HC methods where image blur and complex background induce difficulties for the segmentation step. Performances of DL methods are also limited by image quality as well as the presence of weeds. An hybrid method (HY) was proposed to eliminate weeds between the rows using the rules used for the HC method. HY improves slightly DL performances in the case of high weed infestation. A significant level of variability of plant detection performances is observed between the several experiments. This was explained by the variability of image acquisition conditions including illumination, plant development stage, background complexity and weed infestation. We tested an active learning approach where few images corresponding to the conditions of the testing dataset were complementing the training dataset for DL. Results show a drastic increase of performances for all crops, with relative RMSE below 5% for the estimation of the plant density.
Note: This is a video-only post. The original article body contains only an embedded YouTube video (https://www.youtube.com/watch?v=vaym5Uqshuc) and no text content.
Note: This is a video-only post. The original article body contains only an embedded YouTube video (https://www.youtube.com/watch?v=sHCsOU8byFU) and no text content.
Note: This is a video-only post. The original article body contains only an embedded YouTube video (https://www.youtube.com/watch?v=c_9gQBC9KhU) and no text content.
Note: This is a video-only post. The original article body contains only an embedded YouTube video (https://www.youtube.com/watch?v=u1eyWwqceFo) and no text content.