Method and system for harvesting crops

Abstract

Crop management systems are provided to enable high-throughput harvesting and phenotyping of field crops such as sweet corn. Methods of high-throughput harvesting and phenotyping are also provided that enable market-desirable trait driven genomic selection of plants grown on plots. Plants selected by such systems and methods can be advanced into breeding pipelines and / or advanced for commercialization. These systems and methods enable plants exhibiting high desirable trait phenotype values (such as for marketability traits) to be advanced early.

Description

Technical Field

[0001] The subject matter disclosed herein generally relates to systems, apparatus, and methods for phenotypic analysis and harvesting of field crops, such as maize. In specific embodiments, these systems, apparatus, and methods are suitable for phenotypic analysis and harvesting of crops grown in study plots or experimental plots. Background Technology

[0002] This invention generally relates to a system, apparatus, and method of use that enables high-throughput harvesting and phenotypic analysis of crops, particularly sweet corn, thereby enabling genomic selection of plants driven by market-demand traits. The selected plants can be advanced into breeding pipelines and / or used for commercial purposes.

[0003] Data collected after trials (e.g., in study plots) for phenotypic analysis and selection of sweet corn plants are typically collected at ear maturity (usually 21 days after pollination). At this time, the sweet corn plants are still in an active developmental stage, and the evaluation of harvest data needs to be completed as quickly as possible to prevent the harvest date from becoming a confounding factor. Ideally, the entire site (e.g., the entire plot) should be harvested within one or two days, provided conditions permit.

[0004] To perform the analysis, researchers enter a plot and randomly select a certain number of ears of fruit, a number expected to be representative of the entire plot. In one example, researchers might select five “typical” ears of fruit from a plot and then evaluate all plants in that plot based on an analysis of those five ears (e.g., using an estimated mean). These analyses typically combine objective measurements with subjective assessments.

[0005] The applicant acknowledges that this approach may have potential problems. As a first example, a larger sample size (e.g., 10 or more ears per plot) can provide a more accurate representation of the true mean of the plot. However, obtaining a larger sample size requires more manpower and time. As a second example, some analyses' reliance on subjective assessments may skew the results due to biases in sample selection and estimation.

[0006] Therefore, there is a need to provide higher quality experimental plot data (at least for sweet corn crop analysis), in which sample sizes can be easily scaled up, a balance can be struck between accuracy and labor intensity, and significant trait data can be collected with reduced subjectivity and bias. Summary of the Invention

[0007] According to certain aspects of this disclosure, systems, apparatus, and methods for harvesting crops, such as sweet corn, are disclosed. In specific aspects, these systems, apparatus, and methods enable phenotypic analysis and harvesting of sweet corn crops grown in study or experimental plots. Plants selected using the systems, apparatus, and / or methods disclosed herein can be advanced into a breeding pipeline for commercial purposes. In one example embodiment, improvements to existing harvesting machinery are disclosed that allow for the acquisition of precise optical, quality, and location information during harvesting, while making the equipment transportable and easily deployed in study and experimental networks. Improved data collection from study plots, relying on high-throughput quantitative measurements, provides crop breeders with more reliable data for selection, breeding, and advancement decisions. In an example embodiment, sweet corn crop data collection and analysis can be used to assess marketability traits and advance highly marketable sweet corn plants early in the breeding pipeline. In specific embodiments, the systems, apparatus, and methods disclosed herein allow for automated sample collection and data tracking, for example, through the use of sophisticated GPS and plot barcode information.

[0008] Aspects of the present invention include a harvester configured for harvesting and phenotypic analysis of crops grown in a plot. An exemplary embodiment of the invention relates to a harvester comprising: a harvesting head for gripping a corn plant and removing one or more ears of corn from the gripped corn plant; a collection platform; a collection container operatively coupled to the collection platform for receiving the one or more removed ears of corn; a first sensor configured to sense reception of the one or more removed ears of corn at the collection platform or the collection container; a second geolocation sensor; and a controller including one or more processors communicatively coupled to the first sensor and the second sensor and configurable to execute instructions stored in a computer-readable storage medium. In a particular embodiment, these instructions include: continuously collecting output from the first sensor indicating reception of one or more removed ears of corn from the corn plant at the collection platform or the collection container; and optionally associating the output of the first sensor with the output of the second sensor indicating the location of the corn plant within the plot.

[0009] In an embodiment, the harvester can be further configured to execute instructions including: in response to operator input, discarding one or more removed ears of corn from the collection platform and discarding the outputs of the first and second sensors. Hereinafter, the harvester can operate in a first mode to collect ears of corn while harvesting them from the corn plant, retaining the outputs from the first and second sensors associated with the collected ears; and can operate in a second mode to discard ears of corn while harvesting them from the corn plant, discarding any outputs from the first and second sensors associated with the collected ears.

[0010] In one embodiment, the harvester continuously collects the output from the first sensor, including continuously collecting the first sensor output and extracting the weight of a single corn ear from the output.

[0011] The invention further includes methods of operating a harvester configured for harvesting and phenotypic analysis of crops grown in plots. An example embodiment of this method includes a method for collecting and associating maize information, comprising: harvesting unpeeled maize ears from maize plants in a maize plot; weighing and collecting the weight of the maize ear; collecting the field location of the maize plant; associating the weight of the maize ear with the field location of the maize plant; uniquely identifying the maize ear, the maize plant, and / or the maize plot; imaging the unpeeled maize ear using an imaging device capable of providing an internal image of the maize ear; and extracting phenotypic data of the maize ear from the image. In a particular embodiment, the maize plant is a sweet maize plant. In a particular embodiment, the collection and association are performed continuously as the harvester operates through the maize plot. In a particular embodiment, the collection and association are performed substantially simultaneously.

[0012] In an embodiment, the method further includes associating the phenotypic data of the maize ear with the weight of the maize ear and the field location of the maize plant. In a particular embodiment, uniquely identifying the maize ear, the maize plant, and / or the maize plot includes assigning a unique identifier to the maize ear, the maize plant, and / or the maize plot, and the method further includes associating the phenotypic data of the maize ear with the unique identifier. Aspects of the present invention include an imaging system for phenotypic analysis of crops. Embodiments of such a crop imaging system include: an image capture device; and a processor communicatively coupled to the imaging device, the processor being configured to execute instructions to: capture an image of a crop sample; and extract data associated with traits of the crop sample from the captured image; and, based on the data, assign a separate value to each of one or more attributes associated with the trait of the crop sample. In an embodiment, the processor is further configured to capture images of multiple crop samples harvested from a common plot and to estimate the mean of each attribute associated with the trait of the crop harvested from the common plot based on individual values ​​of each of the one or more attributes for each crop sample from the common plot. In a particular embodiment, the crop is sweet corn. In a particular embodiment, the image capturing device is an imaging device capable of providing an internal image of the corn ear, for example, wherein the image capturing device is an X-ray image capturing device. In other particular embodiments, the image capturing device is any one or a combination of UV, visible light, NIR, CT, or MR image capturing devices, and / or incorporates X-ray imaging capabilities. In a particular embodiment, the trait is a marketability phenotype. In a particular embodiment, the one or more attributes associated with the marketability phenotype include ear stalk length, ear stalk diameter, husk length, ear length, and ear width.

[0013] Various aspects of the present invention also relate to a crop management system capable of harvesting and phenotypic analysis of crops. Example embodiments of the crop management system include: a harvesting subsystem comprising a harvesting head for gripping corn plants and removing one or more unpeeled corn ears from the gripped corn plants; an imaging subsystem comprising an image capture device; one or more sensors including a first sensor for sensing the collection of these corn ears at the harvesting subsystem and a second sensor for sensing the location of the harvesting subsystem; and a processor having code for executing instructions to sense the collection of the removed one or more corn ears at the harvesting subsystem. In embodiments, the crop management system further includes a transfer subsystem for transferring these corn ears from the harvesting subsystem to the imaging subsystem. In an example embodiment, the transfer subsystem includes a conveyor belt.

[0014] Various aspects of the present invention also relate to a method of operating a crop management system, the method including operating a harvesting subsystem and / or an imaging subsystem of the crop management system for harvesting and phenotypic analysis of crops. An example embodiment of a method for collecting plant phenotypes includes: obtaining an image of an unpeeled corn ear containing a corn cob; and extracting significant phenotypic data of the corn cob trait from the image. In an embodiment, the method further includes evaluating the phenotype of the corn cob based on the extracted image. In an embodiment, the method further includes advancing seeds from the corn ear into a breeding pipeline in response to a match between the evaluated phenotype and a desired phenotype. In an embodiment, the method further includes assigning phenotypic values ​​to the corn ear, the corn cob, and / or the corn plant in response to a match between the evaluated phenotype and a desired phenotype. In a particular embodiment, the corn is sweet corn. In a particular embodiment, the phenotype is a marketability trait, and the phenotypic data includes ear length, ear width, or pedicel length. In an embodiment of the method, the image of the unpeeled corn ear is obtained via an imaging subsystem including a camera, an optical sensor, or other known imaging device; and the phenotype of the corn is evaluated via a controller coupled to the imaging subsystem, the controller being configured with code for extracting and / or processing significant phenotypic data of the trait from the image and evaluating the phenotype based on the data.

[0015] Further aspects of the present invention include a method for collecting phenotypes of sweet corn plants. An example embodiment of this method includes: obtaining an image of an unpeeled corn ear containing a cob from a sweet corn plant growing in a plot; extracting data from the image associated with a marketability phenotype of the corn cob; and assigning marketability values ​​to the sweet corn cob, plant, and / or plot based on the data extracted from the image.

[0016] Further aspects of the present invention include a breeding method comprising: harvesting maize ears from maize plants in a plot; obtaining an image of each maize ear; extracting significant phenotypic data of the maize cob from the image of each maize ear; estimating a statistically significant phenotypic value for the maize plants in the plot; and advancing one or more maize plants in the plot into a breeding pipeline based on the statistical value. In a particular embodiment, the maize ears are harvested from maize plants in a test plot and their significant phenotypic characteristics are evaluated to determine whether they should be advanced into a breeding pipeline for commercial purposes. In a particular embodiment, the significant phenotypic characteristic is a marketability phenotypic. In an embodiment, maize plants with statistically significant phenotypic values ​​above a threshold are selected to be advanced into a breeding pipeline. In an example embodiment, the threshold is selected based on a desired significant phenotypic characteristic (e.g., a marketability phenotypic value for the maize cob). Attached Figure Description

[0017] Figure 1 This is a high-level diagram of a crop management system (also referred to herein as "PhenoHarvester") used for high-throughput harvesting and phenotypic analysis of crops (such as sweet corn) grown in plots to enable genomic selection of plants driven by market demand traits, based on the disclosure.

[0018] Figure 2 Images are shown of a precision harvester (also referred to herein as "PhenoPuller") used for harvesting sweet corn from a study plot, in accordance with this disclosure.

[0019] Figure 3 An example embodiment of the PhenoPuller disclosed herein is shown, which includes improvements to existing harvesters (such as the OxBo CP100) to enable the acquisition of higher quality test cell data.

[0020] Figure 4 An example embodiment of PhenoPuller's adjustable RAM holder is shown.

[0021] Figures 5 to 6 An X-ray imaging system (also referred to herein as the "PhenoProcessor") is described, which can be coupled to a precision harvester for acquiring significant trait data from harvested crops with high precision. Figure 5 An example housing of the X-ray imaging equipment and associated support equipment is shown, while Figure 6 An example X-ray imaging apparatus is shown.

[0022] Figure 7 This is an example X-ray image of a sweet corn ear harvested by a harvester. The image is analyzed to extract ear data with significant traits, such as the "marketability" phenotype.

[0023] Figure 8 This is an example X-ray image of a sweet corn ear being analyzed by a harvester for attributes associated with the "marketability" phenotype.

[0024] Figure 9 This is a high-level embodiment of a portable crop imaging system that can be used for high-precision phenotypic analysis of harvested maize ears in the field.

[0025] Figure 10 This is an example implementation of a portable version of Phenoprocessor, which can be used for portable high-speed X-ray phenotypic analysis of harvested maize ears.

[0026] Figure 11 This is an example method for operating a crop management system based on the content of this disclosure. Detailed Implementation

[0027] A system is disclosed that enables high-throughput harvesting of crops in study plots and phenotypic analysis of the harvested crops to collect and analyze trait-significant data. This system allows for trait-driven genomic selection of plants that can be advanced into the breeding pipeline at an early stage. Specific embodiments of the system enable high-throughput harvesting of sweet corn ears in study plots and data collection related to significant sweet corn traits, such as marketability phenotypes. In specific embodiments, the system includes one or more subsystems, one or more devices, or combinations thereof.

[0028] Figure 1 A system 10 is shown for high-throughput harvesting of crops in a study plot and phenotypic analysis of the harvested crops to achieve the collection and analysis of significant trait data. In one embodiment, Figure 1 System 10 is referred to as PhenoHarvester in this paper.

[0029] The PhenoHarvester includes a first subsystem 100 (also referred to herein as "PhenoPuller") for precision harvesting of crops. As detailed below, the PhenoPuller includes one or more means for precision harvesting ears of sweet corn crops in a study plot. In a particular embodiment, the PhenoPuller includes a modification of a sweet corn harvester, adapted to operate as a high-throughput study plot harvester, for example, by adding sensors and a sample collection system.

[0030] The PhenoPuller for precision harvesting can be coupled to a second subsystem 200 (also referred to herein as the "PhenoProcessor") for high-precision image capture of phenotypic data from the harvested crop. In one particular embodiment, as described herein, the PhenoProcessor includes one or more means for high-precision X-ray imaging of the harvested sweet corn ears. In other embodiments, high-precision imaging includes high-precision red-green-blue or RGB imaging, grayscale imaging, ultraviolet or UV imaging, near-infrared or NIR imaging, any other imaging techniques known in the art, and / or combinations thereof.

[0031] PhenoProcessor 200 may further include one or more processors configured to execute instructions stored in a non-transitory computer-readable storage medium to analyze images of harvested crops and extract phenotypic data from those images. In an alternative embodiment, PhenoProcessor may be coupled to a third subsystem including one or more processors for performing image analysis and extracting phenotypic data from images of harvested crops. As used herein, the software module 300 for image analysis is referred to as “SweetTEA” (Sweet Corn Tag Extraction Algorithm). In a particular embodiment, SweetTEA includes computer-readable code implemented via software module 300 to analyze X-ray images of harvested sweet corn ears to obtain data related to marketability phenotypes. As used herein, a marketability phenotype refers to a trait of sweet corn ears related to ear size that enables the packaging of an industry-defined quantity of ears in industry-standard shipping boxes. Marketability traits include several aspects of ear size, including but not limited to ear stalk length, ear stalk diameter, husk length, ear length, and ear width.

[0032] In this way, aspects of the present invention relate to a crop management system capable of harvesting and phenotypic analysis of crops, such as the PhenoHarvester of the present invention. Example embodiments of the crop management system include: a harvesting subsystem comprising a harvesting head for gripping a maize stalk and removing one or more unpeeled maize ears from the gripped stalk; an imaging subsystem comprising an image capture device; one or more sensors including a first sensor for sensing the collection of these maize ears at the harvesting subsystem and a second sensor for sensing the location of the harvesting subsystem; and a processor having code for executing instructions to sense the collection of the removed one or more maize ears at the harvesting subsystem. In embodiments, the crop management system further includes a transfer subsystem for transferring these maize ears from the harvesting subsystem to the imaging subsystem.

[0033] Various aspects of the present invention also relate to a method of operating a crop management system, the method including operating a harvesting subsystem and / or an imaging subsystem of the crop management system to harvest and phenotypic analyze crops (e.g., crops grown in plots). An example embodiment of a method for collecting plant phenotypes includes: obtaining an image of an unpeeled maize ear containing a maize cob; and extracting significant phenotypic data of the maize cob from the image. In an embodiment, the method further includes evaluating the phenotype of the maize cob based on the extracted image. In an embodiment, the method further includes advancing seeds from the maize ear into a breeding pipeline in response to an evaluated phenotype matching a desired phenotype. In a particular embodiment, advancing seeds into the breeding pipeline includes: selecting maize seeds having a desired marketable phenotype, growing maize plants from the selected seeds, and crossing the maize plants including the desired marketable phenotype with another maize plant that does not include the desired marketable phenotype but includes alternative desired agronomic traits (e.g., yield). In an embodiment, the method further includes assigning a phenotypic value to the corn ear, the corn cob, and / or the corn plant in response to an assessed phenotypic match with a desired phenotypic. In a particular embodiment, the corn is sweet corn. In a particular embodiment, the phenotypic is a marketability trait, and the phenotypic data used to assess the marketability trait and assign marketability trait phenotypic values ​​to the corn ear or cob and / or plant includes ear length, ear width, or shank length. In a specific embodiment, the marketability trait of the corn plant is assessed using phenotypic data from a statistically significant number of corn ears harvested from the corn plant, and a marketability trait phenotypic value is assigned to the corn plant. In other specific embodiments, the marketability trait phenotypic value for a given corn plant may be determined based on a weighted or unweighted average, mean, median, or mode of ear length, ear width, and / or shank length data from a statistically significant number (e.g., a threshold number) of corn ears harvested from the corn plant.

[0034] PhenoPuller

[0035] First, the system turns to the precision harvesting subsystem 100 (or PhenoPuller 100), which is responsible for harvesting corn and collecting key yield and yield-related data. In one example embodiment, such as Figures 2 to 4 As shown, PhenoPuller was created by modifying existing commercially available sweet corn (SWC) harvesting equipment into a research plot harvester by improving the machine and adding sensors and a sample collection system. Figures 2 to 4In the illustrated example embodiment of the Phenopuller 100, an improvement upon an existing harvester (OxBo CP100 in the illustrated example) is made to acquire higher quality test plot data. The improved harvester 350 (also referred to herein as Phenopuller 100) can be coupled to a vehicle 352 (e.g., a tractor) via a coupling device 354 (e.g., a traction device), which moves the harvester 350 (or Phenopuller 100) across the field. As the tractor-driven harvester is pulled through plots 330 containing corn stalks, ears are harvested from these stalks and transferred to conveyor belt 356. Thus, there are two basic sweet corn (SWC) markets: a fresh market and a processing market. Ears in the commercial fresh market are typically hand-harvested by teams of workers, then boxed and transported to shops or market stalls for sale. Sweet corn harvested for processing markets is typically done by machines using conventional stripper headers, which are fast and efficient for short-haul transport to processing facilities but can damage the delicate ears of sweet corn in fresh markets. Commercially available mechanical sweet corn harvesters for fresh markets are also available. One example of a manufacturer of such commercially available sweet corn harvesters for fresh markets is the OxBo CP-100 PixAll. Figures 2 to 4 The illustrated Phenopuller 100 embodiment modifies the OxBo CP-100 SWC harvester for use as a research-test harvester. Typically, commercial SWC harvesters include a gripping and cutting system 357 designed to grip the corn stalk above the ear using a pair of moving claw belts 358, cut the stalk below the ear with a rotating disc, and then pull the stalk between a pair of counter-rotating rollers 360 that remove the ear and allow it to fall onto a conveyor belt 356. Empty stalks are discarded beside the machine, and a fan blows away any loose leaves or tassels that may have detached. The ear then falls onto the conveyor belt and accumulates on a rear-mounted trailer 320 towed by the harvester. Depending on the model, the modified harvester 350 (or Phenopuller 100) can have 1 to 8 rows and be either towed or self-propelled.

[0036] The improvements made to the commercially available harvester (to create the Phenopuller 100 and enable research plot harvesting) are described in this paper below.

[0037] Automatic Steering System 390 and its Capabilities

[0038] Conventional harvesters can remove sweet corn ears from plants with minimal damage and minimal debris (leaves / straw). This aspect of the equipment, as well as the tractor 352 used to operate it, remains unchanged. In a particular embodiment, the harvester 350 is improved to include an automatic steering system 390 configured with real-time dynamic geolocation capabilities, providing highly accurate positioning data. In a particular embodiment, the harvester 350 includes an AgLeader Real-Time Dynamic (RTK) Geolocation (GPS) Automatic Steering System 390 coupled to the tractor 352 to assist the operator in focusing on machine operation while keeping the harvester on a predetermined route. The automatic steering system 390 can include location information (e.g., via GPS data) and enables the harvester to automatically steer according to a selected route traversing the plot. This feature of using RTK automatic steering 390 improves accuracy by eliminating errors in satellite signals, such as atmospheric interference and clock inaccuracies. If the test field is planted using automatic steering, the harvester can be easily set to travel in a straight line with inch-level accuracy. This improves the accuracy and efficiency of harvesting ears throughout the plot. Furthermore, because the automatic steering system includes location information, the position of each harvested plant can be determined by referring to its location within the plot. This also makes it easier to identify any plants that may ultimately be selected for advancement in the breeding pipeline. Upper guide belt height adjuster 392

[0039] As part of the harvester setup that can be operated in the field, there is a manual handwheel adjustment device for setting the height of the upper guide belts, which grab the stalks after they are cut and pull them through the machine. Typically, growers only need to harvest one size of corn (usually a single hybrid) in their field, thus requiring only minor adjustments. However, this is not the case in study plots, where hundreds of different hybrids with varying plant heights may be encountered. For this reason, in certain embodiments, the manual handwheel has been replaced with a hydraulic motor, which the driver can adjust directly from the seat in the cab. Based on input received from the user in the tractor cab, the operating height at which the cutting system grips and cuts the stalks at 357 points can be adjusted (e.g., raised or lowered).

[0040] Camera System 394

[0041] The vertical position of the front of the harvester (including the guide belt and cutting system) can be adjusted using the original hydraulic contour wheel system. This allows the machine to conform to the ground contour and makes it easier for the operator to lower the cutting wheels to the optimal position just below the main ear. This allows adjustment of the operating height of the cutting system 357. Typically, corn in a farmer's field is of roughly the same height and of the same hybrid variety, thus requiring little adjustment. However, to achieve good performance in a study plot, adjustments are needed for almost every plot. In a particular embodiment, the harvester system includes a closed-circuit camera system 394 that captures the harvesting view and assists the operator in adjusting both the cutter height and the upper guide belt in a more ergonomic manner. Due to parallax in the cab, it can be difficult to accurately judge the position of these components from the operator's seat inside the tractor 352. By moving the camera system 394 to a more orthogonal position and placing the camera system display more centrally in front of the operator, a better ergonomic position is provided for the operator, allowing for more accurate and repeatable settings for the harvester. The camera system also helps the driver observe blind spots around the machine and the rear operator's workstation (which is obstructed by the machine and cannot be seen by the driver).

[0042] Load transport selector 396

[0043] The standard factory setup for conventional harvesters includes a joystick and a locating pin, used to switch the harvester's operation between a folded "transport" position and an extended "work" position. The default factory method typically requires the operator in the tractor cab to pull back the spring-loaded locating pin on the drawbar, while another person (or more) pushes the cutting system 357 into place to engage the locating pin. This operation can have several drawbacks. Therefore, the applicant has improved this by removing the locating pin and locking system and replacing it with a hydraulic cylinder with a selective locking mechanism. This allows the harvester to be switched between the "work" position (where the cutting system operates to harvest corn ears) and the "transport" position (where the harvester's cutting system is not operating) from the tractor cab 352 without the need for additional ground team involvement. Furthermore, it can be locked in any position between the work and transport positions for optimal tracking of the harvester from behind the tractor.

[0044] Onboard power supply 398

[0045] To provide power to the electronic equipment, the harvester 350 further includes one or more of a portable generator, an air compressor, and a deep-cycle battery with a charger. Each of the above components (if included) is mounted on the harvester. These components provide uninterrupted DC (nominal 12 V) power for operating the electronic equipment and provide compressed air to operate the pneumatic actuators.

[0046] Remote emergency stop device (E-Stop) 364

[0047] Tractor 352 provides power to the harvester via a power take-off (PTO) output (in one embodiment, a John Deere 4066R tractor is used), equipped with an electromechanical PTO engagement system. Improvements to the harvester include modifications to the tractor's wiring so that the PTO can also be started and stopped from the rear operator 362's location. Updated wiring allows the PTO shaft to be stopped from any workstation by pressing emergency stop switch 364, and the PTO shaft will only restart when all switches are in operating mode. This is a highly valuable safety improvement.

[0048] Product conveyor extension

[0049] The main product conveyor was lengthened (by 18 inches in one example embodiment) to provide greater headroom for the ears of fruit to fall into the collection bin 360.

[0050] Rear work platform 370

[0051] The harvester 100 includes a rear working platform 370, which allows the operator 362 to load empty boxes 360, scan barcodes 372 on the boxes, monitor the harvesting process, and remove full boxes from the weighing system 374. The platform is large enough to accommodate multiple turnover boxes / containers and operator workstations 376. The platform is also equipped with safety railings and operator seating.

[0052] 380 portable weighing sensor platform

[0053] As the ears of fruit fall from the end of conveyor 356, they are guided via metal chute 378 into a lightweight plastic bucket or box 360 (e.g., a Uline S-2773B-Y type plastic box) placed on platform 380 of weighing system 374. In a particular embodiment, weighing system 370 is an industrial balance (e.g., a Mettler Toledo balance). The weighing sensor platform 380 of weighing system 374 is situated on a sliding-mounted tray. In embodiments, the sliding-mounted tray allows the platform to be moved to a rear “sample collection” setting where the ears of fruit will fall during sample collection; or to be moved forward to a “bypass” setting where the tray is pulled forward, exposing an opening in the platform floor, allowing the ears of fruit to fall to the ground and back into the plot / field where they were harvested. The “bypass” setting is particularly useful for areas at field boundaries where harvesting is required but no data or samples are needed. This improvement enables the harvesting system (or PhenoPuller) 100 to operate in both a first and a second mode: in the first mode, the ears of fruit are harvested and collected in the collection bin 360, and then guided to the PhenoProcessor for data collection; in the second mode, the ears of fruit are harvested and discarded, and are not guided to the imaging system 200 (see...). Figure 1 (or PhenoProcessor, and does not collect any data for these ear samples.)

[0054] Continuous weighing system 374

[0055] Typical commercial balances based on load cells wait for the load to stabilize before displaying the weight of the object or sending the data to a printer / computer. In a particular embodiment, the harvesting system includes a load cell platform 380 that operates to continuously transmit weight readings of the harvested ears collected in the bin to an onboard computer or processor to record the weight data of the collected ears. The “step” change data generated in the signal at the weighing system can be advantageously used to count the number of ears that fall into the sample bin during the plot harvesting process. The load cell data is then analyzed or processed using software to identify and distinguish between single ear drops and multiple ear drops. In a further embodiment, the load cell data received from the weighing station is additionally used to measure the weight of a single ear and the average weight of an ear. Uniformity in size and weight among ears is a key trait in commercial sweet corn because producers prefer hybrids that produce consistent ear sizes (in length, width, and weight). Uniformity can only be measured by collecting at least a threshold number of ears individually. Ear counts and total weight alone can only provide an average ear weight, but cannot reveal the weight differences between ears. In contrast, the characteristics of continuous streaming weight data collection and analysis enable the assessment of key phenotypic traits.

[0056] Data collection and sample tracking

[0057] During data collection and tracking, operator input is received at the harvesting system to indicate the start and / or end of each plot. This allows for individual tracking of data for each plot and accurate phenotypic assessment of the plants associated with each plot. In a particular embodiment, as the harvester and associated cutting machine move through the field, the operator replaces the sample bins 360 placed on the weighing sensor platform 380 and provides input to the harvesting system's computer indicating the upcoming entry into a new plot. In one embodiment, this can be accomplished using a barcode scanner and sample ID cards located at the operator's station 376. These cards can be pre-printed before harvesting and placed in the intended harvesting order for easy operator access. Once the driver and operator are ready and the harvester has been adjusted and aligned with the next plot to be harvested, the operator scans the correct plot barcode into software that runs on a computer or processor (e.g., a Windows Toughpad) located at the operator's station and mounted on the harvester. This initializes the software to begin collecting a new batch of data, and the driver can begin moving along the plots, heading towards the next path between plots. While the machine operates, the computer records weighing sensor data and precision GNSS signals transmitted by a GPS sensor (e.g., a Trimble SPS986 GPS sensor) located above the cutting head. Location data from the GPS is used to measure the area of ​​the plot from which samples are being collected, and the ears falling from the conveyor into the sample bins allow for the measurement of the individual and total weight of ears that can be harvested from that plot. When the driver observes that the last plant in the plot has been cut by the harvester's cutting system, they can pause to await verbal confirmation from the operator that all ears have passed through the machine and that the bins containing the ears have been removed from the weighing sensors and that new empty bins are in place. The operator then places the barcode card used for the previous plot into a bin along with the ears to maintain consistency with their source plot. The bins containing the ears can then be transported to other data collection stations in the field or back to the research facility for further testing.

[0058] Rear Operator Automation

[0059] In one particular embodiment, the role of the back-end operator is replaced by robotic automation. In embodiments where only ear weight and location data are needed, without the samples themselves, a self-tilting and resetting system can be installed on the weighing sensor platform. For example, this system employs a pneumatic piston and gate setup to push the ears out of the sample box 360, which is fixed to the weighing sensor. This reduces the need for manual resetting of the weighing sensor. With this type of setup, data can also be continuously collected, allowing a computer to determine when to empty the weighing sensor and re-zero it based on weight. A precise yield map of the harvested area can be created without stopping the harvester, where the weight per unit area can be visualized on a display. Additionally, one or more other plot characteristics and attributes (such as population type (e.g., hybrid or inbred line) or agronomic treatments (e.g., fertilizer or biopesticide application)) can be overlaid and used for plant selection based on a combination of multiple factors and phenotypes (e.g., identifying higher-yielding plants).

[0060] Expected data output

[0061] After harvesting, data stored in the processor installed on the harvester can be exported (e.g., exported to an external or remote processor or exported to a cloud processor) and processed to generate a report. This report includes plot attributes such as: plot ID, length of harvested plot (m), total harvest weight (g), number of ears per plot area, average ear weight (g), measure of variation between ears (standard deviation, in g), number of marketable ears (ears meeting specified weight parameters), and marketable yield (in T / ha and the number of boxes containing 48 marketable ears per acre). GIS output describing the plot area vector (e.g., as a JSON file) can be overlaid on UAV orthophoto images or other raster or vector layers, including but not limited to precision seeder seeding data or plant establishment analyzer GIS data.

[0062] In one example embodiment, the PhenoPuller is targeted to harvest 750 single-row plots per day (compared to a typical rate of 150 plots per day for manual harvesting, evaluating a subset of 5 ears per plot). Typically, a two-person team can observe ~1500 ears per day, while the PhenoPuller, with the same two team members, can objectively measure >20,000 ears. In another example embodiment, the PhenoPuller is configured to harvest an entire plot in ~30 seconds and ~100 plots per hour.

[0063] In addition to increased speed, the ability to collect and organize sample boxes for downstream processing is a significant advantage. For example, barcode tags that associate plots with the data being collected can be placed in a sample box along with the collected ears, and the box can be transported to other remote or field-edge processing locations. If sample collection is not required, and only spatial ear count and weight data are desired (e.g., for yield trials), PhenoPuller can be configured to collect data from a plot and then dump the ears onto the ground to prepare for the next plot. This may be particularly beneficial in later stages of product development, where multi-point yield trials are desired and selection of other non-yield traits has been completed. This "collect and dump" mode can be operated by a single operator who both drives the tractor and remotely monitors the weighing sensor system from the cab. PhenoPuller can also be configured with features to improve portability. For example, PhenoPuller can be configured to include a low-bed gooseneck trailer, enabling transport similar to typical agricultural equipment. In one embodiment, the rear platform may also include an adjustable RAM bracket 359 for an onboard processor.

[0064] In this way, commercially available harvesters can be modified to create precision harvesters (or PhenoPullers), which include technological modifications to improve the efficiency of crop data collection in study plots through automation. Therefore, the entire PhenoHarvester system can increase the rate of genetic gain in breeding programs. Specifically, the increased speed and efficiency of obtaining ear counts and weights in single-row study plots facilitates the earlier collection of quantitative yield performance data in the product development pipeline. This not only immediately increases the rate of genetic gain by enabling earlier selection in the hybrid testing pipeline, but also by contributing over time to standardized, robust datasets from which the relationship between yield performance, genotype, and environment can be modeled and used for predictive breeding throughout the product development pipeline.

[0065] Aspects of the present invention include harvesters configured for harvesting and phenotypic analysis of crops grown in plots, such as PhenoPuller. An exemplary embodiment of the invention relates to a harvester comprising: a harvesting head for gripping a corn stalk and removing one or more ears of corn from the gripped stalk; a collection platform; a collection container operatively coupled to the collection platform for receiving the one or more removed ears of corn; a first sensor configured to sense reception of the one or more removed ears of corn at the collection platform or the collection container; a second geolocation sensor; and a controller including one or more processors communicatively coupled to the first and second sensors and configured to execute instructions stored in a computer-readable storage medium. In a particular embodiment, these instructions include: continuously collecting the output from the first sensor, the output indicating reception of one or more removed corn ears from the corn plant at the collection platform or the collection container; and optionally associating the output of the first sensor with the output of a second sensor, the output of which indicates the location of the corn plant within the cell.

[0066] In an embodiment, the harvester can be further configured to execute instructions including: in response to operator input, discarding one or more removed ears of corn from the collection platform and discarding the outputs of the first and second sensors. Hereinafter, the harvester can operate in a first mode to collect ears of corn while harvesting them from the corn plant, retaining the outputs from the first and second sensors associated with the collected ears; and can operate in a second mode to discard ears of corn while harvesting them from the corn plant, discarding any outputs from the first and second sensors associated with the collected ears.

[0067] In one embodiment, the harvester continuously collects the output from the first sensor, including continuously collecting the first sensor output and extracting the weight of a single corn ear from the output.

[0068] The invention further includes methods of operating a harvester configured for harvesting and phenotypic analysis of crops grown in plots. An example embodiment of this method includes a method for collecting and associating maize information, comprising: harvesting unpeeled maize ears from maize plants in a maize plot; weighing and collecting the weight of the maize ear; collecting the field location of the maize plant; associating the weight of the maize ear with the field location of the maize plant; uniquely identifying the maize ear, the maize plant, and / or the maize plot; imaging the unpeeled maize ear using an imaging device capable of providing an internal image of the maize ear; and extracting phenotypic data of the maize ear from the image. In a particular embodiment, the maize plant is a sweet maize plant. In a particular embodiment, the collection and association are performed continuously as the harvester operates through the maize plot. In a particular embodiment, the collection and association are performed substantially simultaneously.

[0069] In one embodiment, the method further includes associating the phenotypic data of the corn ear with the weight of the corn ear and the field location of the corn plant. In a specific embodiment, uniquely identifying the corn ear, the corn plant, and / or the corn plot includes assigning a unique identifier to the corn ear, the corn plant, and / or the corn plot, and the method further includes associating the phenotypic data of the corn ear with the unique identifier.

[0070] PhenoProcessor

[0071] Turn now Figure 1 The PhenoProcessor 200 is configured as a precision imaging subsystem of the acquisition system 10 (or PhenoHarvester). In one example embodiment, such as Figures 5 to 6 As shown, the PhenoProcessor 200 is configured as a portable, high-resolution imaging system 500 for field edge operations. This system is placed near the PhenoPuller in the field, and plot ear samples can be delivered to the system from there (e.g., by members of the harvesting team, or via conveyor belt or other known transfer mechanism). In the depicted embodiment ( Figure 6In this embodiment, the imaging system includes an X-ray imaging system 504 (such as the Loma X-5 SpaceSaver industrial cabinet X-ray system) within a housing 502 (described herein as a trailer). In the depicted embodiment, the imaging system operates as a Field Research Platform (FiRP) or trailer 502, which houses the imaging equipment 504. In addition to the imaging equipment, the FiRP is equipped with support equipment such as a generator (e.g., for providing power to the imaging), an HVAC system (e.g., for maintaining environmental conditions including the temperature at and around the imaging equipment), and a display monitor 602. One or more of these support equipment may be housed inside the housing 502 (e.g., the HVAC system or display monitor 602) or located externally (e.g., the generator 604). Figure 6 The image shows an example view 550 of the interior of the housing or trailer 502, showing the location of the imaging equipment and related support equipment.

[0072] In a particular embodiment, the x-ray imaging system 504 is a dual-beam x-ray imager that provides x-rays of different intensities (e.g., soft x-rays and hard x-rays) to image different parts of the harvested ear of fruit (e.g., to image the filaments of the ear of fruit in a manner different from that of the cob).

[0073] Cellular samples (i.e., ears of fruit harvested by PhenoPuller) are sent to the PhenoProcessor, for example, the harvested ears are sent to FiRP and dumped into containers (e.g., grid carts). Before data collection begins, the barcode label associated with the cellular sample is scanned to establish a link between the cellular sample and the data collected subsequently.

[0074] Imaging settings are selected based on the crop and imaging modality. In one example where the imaging modality is X-ray imaging and the harvested sample is a sweet corn ear, the X-ray settings include: a conveyor speed of 20 m / min, a wavelength between 30 keV and 40 keV, and a current between 1.5 mA and 4.0 mA. The pixel size used can be 0.4 mm in both H and V.

[0075] Each harvested ear of fruit is placed on conveyor belt 606 with a selected common orientation (e.g., base-forward orientation) and is allowed to pass through imaging equipment 504. Ears can be imaged individually or in groups, where individual ears within a group are placed side-by-side for comparison. For example, ears can be imaged individually, in pairs at a time, or in more than two pairs at a time to speed up the process. In one example embodiment, samples can be imaged at a rate of ~2 cells per minute.

[0076] Each ear of fruit (or each pair of ears of fruit, or each group of ears of fruit) is imaged, and the corresponding image is stored in a memory storage medium coupled to the imaging equipment, in a folder labeled according to the associated cell information (information scanned before initiating imaging of the ears of that cell). In one example, each X-ray image of a single ear of fruit is stored as a 256-grayscale BMP image in a folder in computer memory named after the scanned barcode data. Figure 7 An example of such an X-ray image 700 of a harvested maize ear is shown. Acquiring an X-ray image allows for the collection of phenotypic data without removing the husk from the maize ear. For example, X-ray image 700 enables the display of the maize cob region as a shaded area 702 to distinguish it from the husk region (unshaded area 704). Maize cob and husk features can then be collected without prior husking of the ear. This is a non-limiting example and... Figure 8 In detail, X-ray images can be evaluated to segment the ear into the cob region (containing the kernels), the peduncle region located at the rear of the cob region (containing the peduncle through which the ear connects to the corn plant), and the husk region located at the front of the cob region (containing the husks and the silks extending from them). After segmentation, the length, width, diameter, and other parameters of each region can be estimated.

[0077] Once all samples from a given plot (i.e., all harvested ears from that plot) have been imaged, further image acquisition is stopped. Before starting on samples from the next plot, the barcode of the next plot is scanned, opening a new folder associated with that plot in computer memory. Subsequently, all samples associated with that plot are imaged, and the images are stored in the newly created folder.

[0078] After imaging all cells, the image folder can be transferred to a downstream image processing system (e.g., via a USB drive, or via a wired or wireless network connection) to analyze the images and extract salient data from them.

[0079] In this way, aspects of the present invention include imaging systems, such as a Phenoprocessor, for phenotypic analysis of crops. Embodiments of such crop imaging systems include: an image capture device; and a processor communicatively coupled to the imaging device, the processor being configured to execute instructions to: capture an image of a crop sample; extract data associated with a trait of the crop sample from the captured image; and assign a separate value to each of one or more attributes associated with the trait of the crop sample based on the data. In embodiments, the processor is further configured to capture images for multiple crop samples harvested from a common plot and to estimate the average value of each attribute associated with a trait of the crop harvested from the common plot based on the separate value of each of the one or more attributes for each crop sample from the common plot. In a particular embodiment, the crop is sweet corn. In a particular embodiment, the image capture device is an imaging device capable of providing an internal image of the corn ear, for example, wherein the image capture device is an X-ray image capture device. In other particular embodiments, the image capture device is any one or a combination of UV, visible light, NIR, CT, or MR image capture devices, and / or incorporates X-ray imaging capabilities. In a particular embodiment, the trait is a marketability phenotype. In a particular embodiment, the one or more attributes associated with the marketability phenotype include peduncle length, peduncle diameter, husk length, ear length, and ear width.

[0080] SweetTEA (Image Analysis)

[0081] Turning now to SweetTEA, the processing subsystem of PhenoHarvester. In some embodiments, a controller is coupled to PhenoHarvester (e.g., PhenoProcessor) and includes one or more processors configured to execute instructions stored in a non-transitory computer-readable storage medium. These instructions include code for analyzing images of crop samples and extracting phenotypic data of the crop (e.g., data associated with the marketability phenotype of sweet corn ears harvested from a given study plot) from each image of each sample. In some embodiments, the controller is mounted on the PhenoProcessor. In other embodiments, the controller is coupled to the PhenoProcessor but located externally to or remotely from it. In embodiments, the marketability trait extraction algorithm (TEA) is a machine learning (ML)-based software algorithm built into the software package that uses the raw images output by the PhenoProcessor to generate a data output table summarizing objective measurements derived from the images.

[0082] In a particular embodiment, the SweetTEA algorithm is trained using a training dataset that includes images intentionally masked for specific features, enabling the algorithm to learn how to identify these features (e.g., identical or similar features) in unlabeled images of the test dataset. The ML algorithm can be improved over time to identify one or more “standard” features with greater accuracy and to learn new trait salients as they become available. Exemplary trait salients that can be extracted from X-ray images of sweet corn samples and used to evaluate the marketability phenotype of sweet corn include husk length, ear length, ear width, and pedicel length.

[0083] Figure 8 An example of processing an X-ray image 800 of harvested maize ears is shown. Significant phenotypic data (including husk length, ear length, ear width, and pedicel length) are extracted from image 800 without removing the husks from the maize ears. (See previous reference...) Figure 7 As described, the X-ray image 800 enables the corn cob region to be displayed as a shaded area 802, distinguishing it from the husk region (unshaded area 804). Corn cob and husk features can then be collected without pre-peeling the ear. This is a non-limiting example and... Figure 8 In detail, the X-ray image can be evaluated to segment the ear into the cob region (containing the kernels; depicted as the central segment in the image), the peduncle region located at the rear of the cob region (containing the peduncle through which the ear connects to the maize plant) (depicted as the rightmost segment in the image), and the husk region located at the front of the cob region (containing the husks and the silks extending from them) (depicted as the leftmost segment in the image). After segmentation, the length, width, diameter, and other parameters of each region can be estimated. For example, after segmentation, one or more of the following can be determined: peduncle length 816, husk length 812, and ear length 814. Similarly, one or more of the following can be estimated: peduncle or cotyle width 808, ear width 806, and cob width 810.

[0084] Figure 9A high-level example embodiment of a portable imaging system 900 or Phenoprocessor 900 that can be used for portable high-speed X-ray phenotypic analysis of harvested maize ears is shown. The portable imaging system 900 can be configured as a field-side modular platform for use with conventional harvesters. Bags 902 of harvested maize ears 906 are unloaded onto a scale 904 of the portable imaging system 900 for weighing, and a unique identifier (e.g., a barcode) on the bag is scanned. The identifier provides an indication of the maize plant, pedigree, and / or plot from which the ears were harvested and allows the image of the ear to be associated with the identifier of the maize plant, pedigree, and / or plot. In one example embodiment, phenotypic values ​​are evaluated for maize bags from a first plot, followed by processing of maize bags from a different second plot.

[0085] After weighing, the corn ears 906 are then transferred to conveyor 908 for transport to an onboard X-ray imager 910. The corn ears are imaged via X-ray imaging as they pass through the imager 910 on conveyor 908. After imaging, the corn ears 906 are collected in a bin 912 and can be further processed to determine other traits. For example, these ears can then be transferred or moved through another imager (e.g., an NIR, RGB, or UV camera) to collect other phenotypic data or agronomic quality data, such as color, taste, starch content, etc. The data collected from the imaging system can then be used to assign phenotypic values ​​to corn ears associated with unique identifiers. Based on these values, breeders can make breeding selections and decisions.

[0086] Figure 10 Another example embodiment of a portable imaging system 1000 or Phenoprocessor 100, which can be used for portable high-speed X-ray phenotypic analysis of harvested maize ears, is shown. The portable imaging system 1000 can be configured as a ruggedized, integrated device. In a particular embodiment, the portable imaging system 1000 is configured to be dustproof, heat-resistant (e.g., able to withstand at least up to 50°C), and vibration-resistant, and may also be moisture-resistant. The device's dimensions can be designed to be ergonomic, and its overall weight allows for easy transport by a small team or agricultural vehicle. In a particular embodiment, the portable imaging system 1000 weighs 80 lbs and can be carried by two people.

[0087] The harvested corn ears 1002 are received at a first opening 1005 of the portable imaging system 1000. The first opening 1005 is located at a first proximal end of the device and may include a baffle or door that can be closed once imaging is complete. Before being placed for imaging, the bag containing the ears can be weighed, and a unique identifier (e.g., a barcode) 1006 of the bag can be scanned via scanner 1004. The identifier provides an indication of the corn plant, pedigree, and / or plot from which the ears were harvested and allows the image of the ears to be associated with the identifier of the corn plant, pedigree, and / or plot. The ears are placed on a conveyor or moving platform inside the device, thereby moving them to and through the x-ray imager 1008. A shield 1010 protects the operator from x-rays. In one example embodiment, the x-ray imager 1008 captures images at a rate of one ear per second and generates high-resolution images, such as within a 0.2 mm pixel range. In a particular embodiment, the X-ray imager 1008 is a dual-beam X-ray imager that provides both soft and hard X-rays to image the filaments of the ear in a manner different from that of the ear axis. Optionally, the imaging system is communicatively coupled to an onboard or offboard database 1012 for transmitting and storing images (e.g., images can be uploaded to a cloud-connected database). Further, the imaging system is coupled to an onboard or offboard processor 1020 (e.g., a processor in the field or at a workstation) where uploaded images can be processed and phenotypic data can be extracted. Breeder 1022 can make breeding selections and decisions based on the images and the extracted phenotypic data and values. For example, the breeder can analyze the data and make decisions regarding advancement in the pipeline.

[0088] Figure 11 This is an example method 1100 for operating a crop management system according to the present disclosure. The method allows for the collection and imaging of maize ears, and the association of maize information with the collected ears.

[0089] At 1102, the method includes harvesting unpeeled corn ears from corn plants in a corn plot. The unpeeled ears can be harvested by a harvesting system (such as a phenopuller). Alternatively, previously harvested unpeeled ears can be received. In one example embodiment, the ears are received in a bin on the rear platform of the phenopuller.

[0090] At point 1104, determine whether the harvesting system is in collection mode (e.g., based on the position of a joystick or switch, or based on another form of operator input). If not, the system operates in bypass mode, and at point 1130, the collected ears of fruit are discarded without collecting any data from them. For example, the boxes containing the ears of fruit can be placed on a movable platform that can be moved to a location from which the ears of fruit can be easily discarded onto the field floor.

[0091] If the system is in collection mode, at point 1106, the method includes weighing the corn ears and collecting their weight. In a particular embodiment, this collection and the association are performed continuously as the harvester operates through the corn patch. Further, the collection and association are performed substantially simultaneously.

[0092] At 1108, the method includes collecting the field location of maize plants, for example, via GPS sensors coupled to the harvester. At 1110, the method includes associating the weight of maize ears with the field location of maize plants. At 1112, based on this association, the method includes uniquely identifying maize ears, maize plants, and / or maize plots.

[0093] At 1114, the method includes imaging an unpeeled corn ear using an imaging device capable of providing an internal image of the corn ear. For example, X-ray imaging at a phenoprocessor is used to image the unpeeled ear. At 1116, the method includes extracting phenotypic data of the corn ear from the image. In one example, the phenotypic data includes marketability phenotypic data, including ear stalk length, cob length, cob width, etc. Furthermore, phenotypic values ​​associated with marketability can be assigned to the corn ear based on the marketability phenotypic data extracted from the ear image. For example, marketability phenotypic values ​​associated with marketability can be assigned.

[0094] At 1118, the method includes associating phenotypic data of maize ears with the weight of maize ears and the field location of maize plants. In a particular embodiment, uniquely identifying a maize ear, maize plant, and / or maize plot includes assigning a unique identifier to the maize ear, the maize plant, and / or the maize plot, and associating the phenotypic data of the maize ear with the unique identifier.

[0095] At 1120, it is determined whether the phenotypic value (single, absolute, relative, or statistically significant) of the ear, plant, and / or plot is above a threshold. In an example where the phenotypic value is a marketability value, it can be determined whether the maize ear has the desired uniform size. If the phenotypic value is above a threshold (wherein the threshold is selected based on the evaluated phenotypic value), then at 1122, the associated maize plant is selected to advance into the breeding pipeline, or an alternative favorable breeding association decision is made. Furthermore, other plants from the associated plot can be considered favorable or indicated as requiring further evaluation for other phenotypes.

[0096] If the phenotypic association value is below the threshold, at point 1124, the associated maize plant is not selected to proceed into the breeding pipeline, or an alternative unfavorable breeding association decision is made. The associated plant can be discarded. Furthermore, other plants from the associated plot can be considered unfavorable or indicated as not requiring further evaluation for other phenotypes. In this way, the PhenoHarvester system has the potential to improve the genetic gain rate of crop breeding programs (particularly sweet maize breeding programs) by automating data collection efficiency. Specifically, the increased speed and efficiency of obtaining ear counts and weights in single-row study plots will facilitate the earlier collection of quantitative yield performance data in the product development pipeline. This will not only immediately improve the genetic gain rate by enabling earlier selection in the hybrid testing pipeline, but also by contributing over time to standardized, robust datasets from which the relationship between yield performance, genotype, and environment can be modeled and used for predictive breeding throughout the product development pipeline.

[0097] In this way, aspects of the present invention include a method for collecting phenotypes of sweet corn plants. An example embodiment of this method includes: obtaining an image of an unpeeled corn ear containing a cob from a sweet corn plant growing in a plot; extracting data from the image associated with a marketability phenotype of the corn cob; and assigning a marketability value to the sweet corn cob, plant, and / or plot based on the data extracted from the image.

[0098] The invention further includes breeding methods. An example embodiment of this method includes: harvesting maize ears from maize plants in a plot; obtaining an image of each maize ear; extracting significant phenotypic data of the maize cob from the image of each maize ear; estimating a statistically significant phenotypic value for the maize plants in the plot; and, based on the statistical value, advancing one or more maize plants from the plot into a breeding pipeline. In a particular embodiment, the significant phenotypic trait is a marketability phenotypic.

[0099] In a further embodiment, the crop management system or PhenoHarvester 10 may include one or more of the following features:

[0100] 1. Straw detection in PhenoPuller serves as an additional or alternative population calculation tool. In a particular embodiment, a stalk detection "finger" (directly below the Trimble GPS detector) is added to PhenoPuller directly below and behind the straw cutter. This finger attachment sends a straw detection trigger signal in response to gripping the straw, thereby providing an indication of the actual count of straw harvested within the cell area. This can be used to understand harvesting populations and spacing issues.

[0101] 2. PhenoPuller GIS data layer analysis helps understand the exact harvested area (e.g., with an accuracy of 5 cm) for creating a given harvester sample (e.g., the location of the plant harvesting a given ear of fruit). Understanding this allows for a highly accurate estimate of fresh weight yield (Mg / ha). Combined with PhenoProcessor data, a highly accurate measure of marketable ears per Ha can be obtained.

[0102] 3. PhenoPicker as an improvement on PhenoPuller. In this paper, a stripper header can be used instead of a harvester to remove the ear stalks of corn, thus enabling its use in processing sweet corn or commercial dent corn.

[0103] 4. Utilize emergence data to identify improved germination traits. Germination and early seedling development data are important traits in sweet corn because endosperm mutations in sweet corn can restrict starch development, leading to insufficient energy supply in new seedlings. By combining precise seed drop date data, seedling emergence data from plant emergence analyzers or UAVs, and PhenoPuller stem detection data, the controller can be coded to determine the extent to which seeds germinate and ultimately develop into functional plants.

[0104] 5. Downstream phenotypic analysis from PhenoProcessor can be used, including yellow camera and / or sensory testing. Since X-rays are a non-destructive sensor system, the ears of fruit can be transferred to another sensor or imaging system for further analysis after measurement. One example is the use of a camera-based system (yellow camera) that can image subsamples of peeled ears to measure kernel color or other visual traits.

[0105] 6. Phenoprocessor X-ray analysis for starch content. In certain embodiments, the X-ray system can be used to detect or measure the relationship between sugar (sucrose) water and starch in situ, because the X-ray absorption rate of starch crystals differs from that of sugar in solution. Since starch formation in sweet corn ears is undesirable, a high-throughput, non-destructive method for starch detection can be developed. This method allows the same X-ray data from corn ears to be used for both marketability phenotypic correlation data and sugar content estimation.

[0106] Many modifications and other implementations of this disclosure will occur to those skilled in the art, taking advantage of the teachings presented in the foregoing description and associated drawings. Therefore, it should be understood that this disclosure is not limited to the specific embodiments disclosed herein, and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terminology is used herein, it is used only in a general and descriptive sense and not for limiting purposes.

Claims

1. A harvester, comprising: A harvesting head, used to clamp a corn plant and remove one or more ears of corn from the clamped corn plant; Collection platform; A collection container operatively connected to the collection platform, the collection container being used to receive the one or more removed corn ears; A first sensor is configured to sense the reception of the one or more removed corn ears at the collection platform or the collection container. Second geolocation sensor; The controller includes one or more processors communicatively coupled to the first sensor and the second sensor and configured to execute instructions stored in a computer-readable storage medium, the instructions including: The output from the first sensor is continuously collected, the output indicating the reception of one or more removed corn ears from the corn plant at the collection platform or the collection container; as well as Optionally, the output of the first sensor can be associated with the output of the second sensor, the output of the second sensor indicating the location of the corn plant within the plot.

2. The harvester of claim 1 is further configured to execute instructions including: in response to operator input, discarding one or more removed ears of grain from the collection platform and discarding the outputs of the first sensor and the second sensor.

3. The harvester as described in claim 1, wherein, Continuously collecting the output from the first sensor includes: continuously collecting the first sensor output and extracting the weight of a single corn ear from the output.

4. A method for collecting and associating corn information, the method comprising: Harvest unpeeled corn ears from corn plants in the cornfield; The corn ears were weighed and their weights collected. Collect the field locations of the corn plants; The weight of the corn ear is correlated with the field location of the corn plant; The corn ears, the corn plants, and / or the corn plots shall be uniquely identified; Imaging the unpeeled corn cob using an imaging device capable of providing an internal image of the corn cob; and Extract the phenotypic data of the corn ears from the image.

5. The method of claim 4, wherein, The corn plant in question is a sweet corn plant.

6. The method of claim 4, wherein, The collection and association are performed continuously as the harvester moves through the cornfield.

7. The method of claim 6, wherein, The collection and association are performed substantially simultaneously.

8. The method of claim 4, further comprising correlating the phenotypic data of the corn ear with the weight of the corn ear and the field location of the corn plant.

9. The method of claim 4 or 8, wherein, Uniquely identifying the corn ear, the corn plant, and / or the corn plot includes assigning a unique identifier to the corn ear, the corn plant, and / or the corn plot, and the method further includes associating the phenotypic data of the corn ear with the unique identifier.

10. A crop imaging system, comprising: Image capture device; as well as A processor communicatively coupled to the imaging device, the processor being configured to execute instructions to: Capture images of crop samples; as well as Extract data associated with traits of the crop sample from the captured images; as well as Based on the data, a separate value is assigned to each of one or more attributes associated with the trait of the crop sample.

11. The system of claim 10, wherein, The processor is further configured to: Images are captured for multiple crop samples collected from a common plot, and the mean value of each attribute associated with the trait of the crop collected from the common plot is estimated based on the individual value of each of the one or more attributes of each crop sample from the common plot.

12. The system as claimed in claim 10 or 11, wherein, The crop in question is sweet corn.

13. The system of claim 10, wherein, An image capture device is an imaging device capable of providing an image of the interior of the corn cob.

14. The system of claim 13, wherein, The image capture device is an X-ray image capture device.

15. The system of claim 10, wherein, The image capture device is any one of UV, visible light, NIR, CT or MR image capture devices, and wherein the system is portable.

16. The system of claim 10, wherein, The traits described are marketable phenotypes.

17. The system of claim 16, wherein, The one or more attributes associated with the marketability phenotype include ear stalk length, ear stalk diameter, husk length, ear length, and ear width.

18. A crop management system, comprising: A harvesting subsystem, comprising a harvesting head for gripping corn stalks and removing one or more unpeeled corn ears from the gripped corn stalks; An imaging subsystem, comprising an image capture device; One or more sensors, including a first sensor for sensing the collection of the corn ears at the harvesting subsystem and a second sensor for sensing the location of the harvesting subsystem; as well as A processor having code for executing instructions to sense the collection of one or more removed corn ears in the harvesting subsystem.

19. The system of claim 18, further comprising a transfer subsystem for transferring the corn ears from the harvesting subsystem to the imaging subsystem.

20. A method for collecting plant phenotypes, the method comprising: Obtain an image of an unpeeled corn cob containing the corn cob; as well as Extract significant phenotypic data of the maize cob traits from the image.

21. The method of claim 20, further comprising evaluating the phenotype of the maize cob based on the extracted image.

22. The method of claim 21, further comprising advancing seeds from the maize ear into a breeding pipeline in response to an assessed phenotype matching a desired phenotype.

23. The method of claim 21, further comprising assigning phenotypic values ​​to the maize ear, the maize cob, and / or the maize plant in response to an assessed phenotypic matching a desired phenotypic.

24. The method of claim 21, wherein, The corn in question is sweet corn.

25. The method of claim 24, wherein, The phenotype is a marketability trait.

26. The method of claim 25, wherein, The marketability traits are ear length, ear width, or ear stalk length.

27. A method for collecting phenotypes of sweet corn plants, the method comprising: Obtain images of unpeeled corn ears, including the cob, from sweet corn plants growing in the community; as well as Extract data associated with the marketability phenotype of the corn cob from the image; as well as Based on the data extracted from the image, marketability values ​​are assigned to the sweet corn cobs, plants, and / or plots.

28. A breeding method, comprising: Harvest corn ears from the corn plants in the community; Obtain an image of each corn cob; Significant phenotypic data of the corn cob traits are extracted from the images of each corn ear; Estimate the statistically significant phenotypic values ​​of the maize plants in the aforementioned plot; as well as Based on the statistical values, one or more corn plants from the plot are introduced into the breeding pipeline.

29. The method of claim 28, wherein, The significant phenotype of the trait is the marketability phenotype.

Images