AUTOMATED OUTDOR PHENOTYPING
The Field Scanalyzer Gantry System is a 3-axis sensor-to-plant phenotyping system. The design and construction bases on an industrial portal crane system. The x-axis is guided along a rail system underpinned by concrete piles driven into the ground so as to allow natural drainage and no impediments as traditional concrete footings may act as flow barriers. In x-direction, length is only limited by the cabling requirements; one of our customer installations reaches 500 m length. The y axis is orthogonal to the rails and bears the lifting unit for the container with the sensing equipment. In y-direction customised width, e.g. 10 m, 20 m or 30 m is possible to span over a given growth area. The z axis serves to lift up and down the container with the sensor equipment.
This setup will allow a precise movement of all 3 axes throughout the year. The sensor box can reach each point of the measuring area in pre-defined schedules with repeatable high precision in the sub-centimetre range.
The system carries a sensor box for phenotyping sensors and environmental sensors. Any of the camera system will be housed in a separate weather sealed housing and mounted on a flexible and adjustable platform within the imaging box. The total payload of the camera box can reach up to 500kg allowing full flexibility for future adjustments and expansions of the sensor platform. The weather shielding of the sensor box protects the electronic equipment during harsh weather. This comprises a roller door at the down facing side of the platform to fully protect the sensors while not in use.
Having environmental sensors on board, the Field Scanalyzer records climatic data during all phenotypic measurements so that users find phenotype and environment data linked in the database.
The Field Scanalyzer Sensor Box can harbour the following sensor systems:
|Field Scanalyzer Sensor to Plant - sensor options|
|Visible-light (VIS) Camera Module||Size, count, colour, morphology, texture, movement|
|Near infrared (NIR) Camera Module||Reflectance in the water band at 1450 nm|
|Chlorophyll Fluorescence Kinetics Module||Fluorescence induction imaging, chlorophyll status and activity|
|Hyperspectral Imaging Module||Spectrally resolved reflectance|
|Multispectral Imaging Module||Reflectance at a series of distinct wavelengths|
|Infrared (IR) Camera Module||Thermal radiation in the range of 7500 - 13000 nm|
|3D Laser Scanning Module||3D point cloud, height and angle|
|Sensors for weather and environmental conditions||Temperature, light, air humidity, CO2|
Sadeghi-Tehran P, Sabermanesh K, Virlet N, Hawkesford MJ (2017) Automated Method to Determine Two Critical Growth Stages of Wheat: Heading and Flowering. Front. Plant Sci. 8:252. http://journal.frontiersin.org/article/10.3389/fpls.2017.00252/full
Sadeghi-Tehran P, Virlet N, Sabermanesh K, Hawkesford MJ (2017) Multi-feature machine learning model for automatic segmentation of green fractional vegetation cover for high-throughput field phenotyping. Plant methods 13:422
Virlet N, Sabermanesh K, Sadeghi-Tehran P, Hawkesford MJ (2017) Field Scanalyzer: An automated robotic field phenotyping platform for detailed crop monitoring. Functional Plant Biology 44:143
All sensors are mounted in a separate weather-proof cabinet which is moved over the field by the supporting gantry. The maximum total payload of the sensor bay is 500 kg which allows for additional sensors if required.
The sensor bay contains a range of cameras together with dedicated illumination devices. It also contains equipment for sensing environmental data.
Visible light imaging delivers photographs that can be processed for dimensions, morphology, and colour of the samples. They are widely used to monitor growth, development, or environmental responses of plants and canopies.
Infrared imaging delivers images representing the heat emission from the sample surface. Such images can serve to measure canopy temperatures and can be related to transpirational processes.
Hyperspectral imaging, also known as imaging spectrometry, is widely used in remote sensing. The FIELD SCANALYZER is able to image at close range (2.2 m above the canopy) and with high repeatability. Hyperspectral cameras produce a stack of images (hyperspectral data cube), where each image represents a narrow wavelength range of the electromagnetic spectrum.
PS2 Fluorescence analysis addresses status and functions of the chlorophylls that are main components of the light harvesting complexes. The system emits an intense LED flash for 1 second during which 25 images are recorded. From these images variable fluorescence (Fv) can be compared to the saturation level of fluorescence (Fm) and used to assess quantum yield of the PS2 photochemistry.
Laser Scanners are able to scan the plant canopy with very high cubicle resolution using an NIR Laser (840 nm) to ensure high reflectance by plant tissue and minimal physiologic interaction (eg chlorophyll excitation). Below is an example of coarse synthetic point cloud for field data, visualized with white points using shading for a better characterization. A height map can be extracted by encoding the height with colour.
As phenotypes strongly depend on environmental influences, particularly when plants grow in a field, environmental sensors are important to deliver data on factors such as temperature or light. These environmental data are stored together with the camera measurements.
|Sensor options||Visible light camera, 8.1 Megapixel
Chlorophyll fluorescence camera, 1.4 Megapixel
Infrared camera, 0.3 Megapixel
Hyperspectral cameras 380 – 1000 nm; 900 – 2500 nm
3D Laser scanner, 0.25 mm resolutionEnvironmental sensors
|Addressable crop area (examples)||10 m x 110 m
20 m x 200 m
|Control||Dedicated Windows PC plus database server(s)|
|Software options||Process control, Image recording and processing, Data analysis|
WHEAT EARS SEGMENTATION
This process may be used to propose the development stage of the wheat by analysing the size of the ears.