Removal of noise from geographical raster datasets

Abstract

Systems, methods, and apparatuses for the selection of high yield crop varieties are described. Raw raster data comprising points associated with yield values in portions of a geographic area may be acquired. Points that are statistical outliers may be removed from the raster data and used to determine regions of the raster data that comprise fewer points than a threshold number of the plurality of points. Spatial outliers may be detected and a raster variogram from which the spatial outliers have been removed may be generated. Clean raster data may be generated based on the raster diagram and application of kriging techniques to the regions of the raster data which the spatial outliers have been removed. Based on the clean raster data, types and amounts of crops to cultivate in each of the portions of the geographic area may be selected.

Description

REMOVAL OF NOISE FROM GEOGRAPHICAL RASTER DATASETSTECHNICAL FIELD OF USE

[0001] Aspects of the disclosure relate to the automatic removal of noise from geographical datasets used in selection of crop varieties. Aspects of the disclosure provide for using a mechanized yield monitor to determine crop yield from which raster data may be generated and then cleaned for use in selecting high yield crop varieties. Additional aspects provide for more accurately correlating crop yield with crop management practices.BACKGROUND

[0002] Crop breeding decisions may rely on various forms of data including data that indicates the state of crops in an area. For example, crop breeding decisions may rely on images of fields in which crops are grown, that may be analyzed over time to determine various information about the crops including yield. In some cases, analysis of the data may require a significant amount of manual intervention which may increase the processing costs, both in terms of time and money. Further, raw data may comprise a significant amount of noise that may reduce the usefulness of the data and potentially lead to incorrect decisions based on the noisy data. Accordingly, there may be challenges in improving the ways in which data may be processed.SUMMARY

[0003] The following presents a simplified summary of various aspects described herein. This summary is not an extensive overview and is not intended to identify key or critical elements or to delineate the scope of the claims. The following summary merely presents some concepts in a simplified form as an introductory prelude to the more detailed description provided below.

[0004] Aspects described herein may allow for automatic methods, systems, non-transitory machine-readable media, and / or devices for the automated removal of noise from geographical data for use in a crop breeding program. More particularly, some aspects described herein may provide a system for selecting the highest yielding crop variety in a geographic area. The system may comprise a mechanized yield monitor configured to determine a plurality of yield values of one or more crops within a plurality of portions of a geographic area, and generateraw raster data comprising a plurality of points corresponding to the plurality of portions of the geographic area. The plurality of points may be associated with the plurality of yield values in the plurality of portions of the geographic area. The system may comprise a computing device configured to receive the raw raster data. Such raw raster data may be data gathered by the sensor directly in raster form, or collected as a combination of sensor signals aggregated into time segments and represented as polygons, which polygons represent the area covered during the time interval, with the raster data being the values distributed across an underlying raster grid of cells to which the polygons intersect and / or overlap. The computing device may remove, from the raw raster data, a plurality of points that meet one or more statistical outlier criteria. The computing device may determine, based on application of one or more spatial outlier criteria to the raw raster data from which the statistical outliers have been removed, a threshold point density of the plurality of points to use in detection of spatial outliers. The computing device may determine a plurality of regions based on the raw raster data from which the statistical outliers have been removed. Each of the plurality of regions may comprise a different set of the plurality of points, and wherein each of the plurality of regions has a lower point density than the threshold point density. The computing device may recursively remove, from the plurality of regions, based on application of spatial outlier detection techniques to each of the plurality of regions, the plurality of points that meet one or more spatial outlier criteria. The spatial outlier detection techniques may comprise detecting spatial outliers in the plurality of regions, with a number and size of regions determined by recursive division of coordinate space to meet a maximum size threshold. The spatial outlier detection techniques may comprise spatial clustering analysis of the plurality of points in each of the plurality of regions. The computing device may generate a raster variogram based on the plurality of regions from which the spatial outliers have been removed. The computing device may generate, based on the raster variogram and application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data. The computing device may select, based on the clean raster data, the crop variety that is highest yielding for the geographic area. The computing device may generate an indication for a plant breeding program to advance development of the crop variety that is the highest yielding. The crop variety may be an inbred variety, a hybrid variety, or a varietal.

[0005] According to some aspects described herein, the system may be further configured to recommend the selected highest yielding crop variety for planting in a location withenvironmental characteristics that match the geographic area in which the variety was evaluated.

[0006] According to some aspects described herein, the recommending may comprise recommending a multi-hybrid planter, or multi-variety planter, to plant seeds using crop cultivation instructions.

[0007] According to some aspects described herein, the system may be further configured to generate an image of the geographic area based on the clean raster data. The image may comprise a plurality of indications based on the plurality of yield values associated with the plurality of portions of the geographic area.

[0008] According to some aspects described herein, the one or more crops comprise at least two crop varieties. Further, the plurality of indications indicate which of the at least two crop varieties may be planted in each of the plurality of portions of the geographic area. The crop varieties may be inbred varieties, hybrid varieties, or varietals.

[0009] According to some aspects described herein, the method comprises determining the optimal management practices to utilize with a particular variety.

[0010] According to some aspects described herein, the method comprises determining the optimal variety to match with a grower’s preferred management practices.

[0011] Some aspects described herein may provide a method of selecting and planting a crop based on raster data. The method may comprise acquiring, by a computing device comprising one or more processors, raw raster data comprising a plurality of points corresponding to a plurality of portions of a geographic area. The plurality of points may be associated with a plurality of yield values in the plurality of portions of the geographic area. The method may comprise, based on detection of a plurality of points that may be statistical outliers, removing, by the computing device, from the raw raster data, the plurality of points that may be statistical outliers. The method may comprise, determining, by the computing device, based on application of one or more statistical outlier criteria to the raw raster data, a plurality of regions based on the raw raster data from which the statistical outliers have been removed. Each of the plurality of regions may comprise a different set of the plurality of points, and each of the plurality of regions may comprise fewer points than a threshold number of the plurality of points. The method may comprise, recursively detecting, by the computing device,based on application of one or more spatial outlier detection techniques to each of the plurality of regions, spatial outliers in the plurality of regions. The spatial outlier detection techniques comprise spatial clustering analysis of the plurality of points in each of the plurality of regions. The method may comprise, generating, by the computing device, a raster variogram based on the plurality of regions from which the spatial outliers have been removed. The method may comprise, generating, by the computing device, based on the raster variogram and application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data. The method may comprise, selecting, by the computing device, based on the clean raster data, a plurality of amounts of the one or more crops to cultivate in each of the plurality of portions of the geographic area. The method may comprise, generating, by the computing device, based on the clean raster data, crop cultivation instructions based on the plurality of amounts of the one or more crops to cultivate in each of the plurality of portions.

[0012] According to some aspects described herein, the method may further comprise correlating, by the computing device, the crop cultivation instructions with historical crop growth factors for the geographic area, the historical crop growth factors for the geographic area comprising daily precipitation, daily evapotranspiration, irrigation practices, or soil type; and determining, by the computing device, based on the correlation between the crop cultivation instructions and the historical crop growth factors, the one or more crops to cultivate in each of the plurality of portions of the geographic area.

[0013] According to some aspects described herein, the method may further comprise providing the crop cultivation instructions to one or more crop cultivation devices configured to cultivate the one or more crops based on the crop cultivation instructions.

[0014] According to some aspects described herein, the crop cultivation instructions comprise plant density prescriptions.

[0015] According to some aspects described herein, the determining the plurality of amounts of the one or more crops to cultivate is positively correlated with the plurality of yield values in the respective plurality of portions of the geographic area.

[0016] According to some aspects described herein, the one or more statistical outlier criteria comprise a minimum yield value or a maximum yield value. Further, the minimumyield value or the maximum yield value may be based on a standard deviation of the plurality of yield values.

[0017] According to some aspects described herein, the spatial outlier detection techniques may be performed serially on one of the plurality of regions at a time.

[0018] According to some aspects described herein, the spatial outlier detection techniques may be performed in parallel on more than one of the plurality of regions at a time.

[0019] According to some aspects described herein, the one or more spatial outlier detection techniques comprise determining Moran’s I for the plurality of points in the plurality of nonoverlapping regions.

[0020] According to some aspects described herein, the one or more spatial outlier detection techniques comprise determining a plurality of Cook’s distances based on the plurality of yield values associated with the plurality of points in the plurality of regions.

[0021] According to some aspects described herein, the spatial outliers comprise the plurality of points corresponding to the plurality of Cook’s distances greater than a predetermined Cook’s distance threshold.

[0022] According to some aspects described herein, the method may comprise determining an amount of computational resources that may be available, and determining the threshold number of the plurality of points to analyze recursively based on the amount of computational resources available.

[0023] According to some aspects described herein, the plurality of points in each of the plurality of regions may be mutually exclusive.

[0024] According to some aspects described herein, the one or more crop cultivation devices comprise a multi-hybrid or multi-variety planter and / or an irrigation device.

[0025] Some aspects described herein may provide a method of selecting and planting a crop based on raster data. The method may comprise acquiring, by a computing device comprising one or more processors, raw raster data comprising a plurality of points corresponding to a plurality of portions of a geographic area. The plurality of points may be associated with a plurality of yield values in the plurality of portions of the geographic area.The method may comprise, determining, by the computing device, based on removal of statistical outliers from the raw raster data, a plurality of regions comprising fewer points than a threshold number of the plurality of points. The method may comprise, detecting, by the computing device, based on application of spatial outlier detection techniques to each of the plurality of regions, spatial outliers in the plurality of regions. The spatial outlier detection techniques comprise spatial clustering analysis of the plurality of points in each of the plurality of regions. The method may comprise, generating, by the computing device, a raster variogram based on the plurality of regions from which the spatial outliers have been removed. The method may comprise, generating, by the computing device, based on the raster variogram and application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data.

[0026] According to some aspects described herein, the one or more crops comprise at least two hybrid crop varieties, and wherein the plurality of indications indicate which of the at least two hybrid crop varieties is planted in each of the plurality of portions of the geographic area.

[0027] According to some aspects described herein, the one or more spatial outlier detection techniques may comprise determining a plurality of Cook’s distances based on the plurality of geospatial values associated with the plurality of points in the plurality of regions.

[0028] According to some aspects described herein, the spatial outliers may comprise the plurality of points corresponding to the plurality of Cook’s distances greater than a predetermined Cook’s distance threshold.

[0029] According to some aspects described herein, the plurality of geospatial values may comprise yield values that were obtained from an impact flow sensor.

[0030] According to some aspects described herein, the plurality of geospatial values may comprise yield values that were obtained from a split planter trial.

[0031] According to some aspects described herein, the cleaned raster data may be utilized to select agronomic practices to utilize for a selected variety in a geographic area.

[0032] According to some aspects described herein, the plurality of geospatial values may be used as planted seed planting density values.

[0033] According to some aspects described herein, the plurality of geospatial values may be used as applied values of a crop input.

[0034] Corresponding apparatuses, devices, systems, and / or computer-readable media (e.g., non-transitory computer readable media) are also within the scope of the disclosure. By generating cleaned raster data with reduced noise crop yield may be more accurately determined and the performance of a crop breeding program may be improved. These features, along with many others, are discussed in greater detail below.BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present disclosure is illustrated by way of example and not limited in the accompanying figures in which like reference numerals indicate similar elements and in which:

[0036] FIG. 1 shows an example of a computing system that may be used to implement one or more aspects of the disclosure in accordance with one or more illustrative aspects discussed herein.

[0037] FIG. 2 shows an example of a raw yield data, cleaned data, and yield difference data according to one or more aspects of the disclosure.

[0038] FIG. 3 shows a timing diagram that indicates a data flow of signals associated with removal of noise from geographical raster datasets according to one or more aspects of the disclosure.

[0039] FIG. 4 shows an example flow chart for raster cleaning and error estimation of spatial data according to one or more aspects of the disclosure.

[0040] FIG. 5 shows an example flow chart for spatial outlier detection according to one or more aspects of the disclosure.

[0041] FIG. 6 shows an example flow chart for error estimation according to one or more aspects of the disclosure.

[0042] FIG. 7 shows an example flow chart for raster cleaning according to one or more aspects of the disclosure.

[0043] FIG. 8 shows an example flow chart for use of cultivation instructions according to one or more aspects of the disclosure.DETAILED DESCRIPTION

[0044] In the following description of the various embodiments, reference is made to the accompanying drawings, which form a part hereof, and in which is shown by way of illustration various embodiments in which aspects of the disclosure may be practiced. It is to be understood that other embodiments may be utilized and structural and functional modifications may be made without departing from the scope of the present disclosure. Aspects of the disclosure are capable of other embodiments and of being practiced or being carried out in various ways. Also, it is to be understood that the phraseology and terminology used herein are for the purpose of description and should not be regarded as limiting. Rather, the phrases and terms used herein are to be given their broadest interpretation and meaning. The use of “including” and “comprising” and variations thereof is meant to encompass the items listed thereafter and equivalents thereof as well as additional items and equivalents thereof.

[0045] Aspects described herein are generally directed to the removal of noise from geographical raster datasets. Generating cleaned geographical raster datasets that have had noise removed or reduced may improve the ability to select crop varieties with higher yield. The disclosed technology may be used to generate noise-reduced geographical raster data that may be used to improve the outcomes of a plant breeding program. The noise-reduced geographical data may allow for more effective determination of high yield geographic areas that may produce improved crop yields.

[0046] The disclosed technology may leverage the use of novel techniques to detect spatial and statistical outliers in data, which may be applied to raster data and used to generate cleaned raster data that may be used in a crop breeding program or to select varieties for commercial production in a specific environment. According to the aspects described herein, these and other technical effects and benefits may be achieved through the use of cleaned raster data to generate cultivation instructions that may be used by cultivation devices (e.g., planting devices) to more effectively cultivate crops.

[0047] Before discussing these concepts in greater detail, however, several examples of computing devices and / or computing systems that may be used in implementing and / orotherwise providing various aspects of the disclosure will first be discussed with respect to FIG. 1.

[0048] FIG. 1 shows an example of a computing system that may be used to implement one or more aspects of the disclosure in accordance with one or more illustrative aspects discussed herein. In particular, FIG. 1 depicts a diagram of a computing system that may be configured to perform operations comprising the exchange and processing of signals and / or data that may be used to reduce noise in geographical raster data. As shown in FIG. 1, the computing system includes the network 102, a computing system 104, a yield monitor device 106, a multi -hybrid planter 108, and an irrigation system 110. Computing system 104 may comprise one or more processors 112, memory 114, raw raster data 120, cleaned raster data 122, and / or crop cultivation instructions 124. System 100 may operate in a standalone environment and / or as part of a networked environment that may include other devices and / or systems. For example, system 100 may operate in conjunction with other computing systems and / or other computing devices not shown in FIG. 1. As shown in FIG. 1, various computing devices including computing system 104, yield monitor device 106, multi -hybrid planter 108, and / or irrigation system 110 may be interconnected via the network 102. Further, the system 100 may operate via one or more networks not including the network 102 shown in FIG. 1. For non-hybrid crops, a multi-variety planter would be used for 108.

[0049] Network 102 may be used to communicate (e.g., send and / or receive) signals, information, and / or data. For example, network 102 may be used to communicate crop cultivation instructions 124 that may be sent from computing system 104 to multi-hybrid planter 108. Network 102 may include any combination of wired and / or wireless networks and may carry any type of signal or communication including communications and / or signals using one or more communication protocols (e.g., TCP / IP, HTTP, and / or HTTPS). Further, network 102 may include any combination of a local area network (LAN), an intranet, a wide area network (WAN), and / or the Internet. Furthermore, network 102 may be configured or arranged according to any known topology and / or architecture.

[0050] Computing system 104 may, in some embodiments, implement one or more aspects of the present disclosure by accessing and / or executing instructions; and / or performing one or more operations based at least in part on the instructions. For example, computing system 104 may generate instructions (e.g., crop cultivation instructions 124) that may comprise crop cultivation instructions that may be used by the multi -hybrid planter 108 to plant seeds. In someembodiments, the system 100 may be incorporated into and / or include a computing device (e.g., a computing device with one or more processors, one or more memory devices, one or more input devices, and / or one or more output devices). For example, computing system 104 may be incorporated into and / or include a desktop computer, a computer server, a computer client, a mobile device (e.g., a laptop computer, a tablet computer, a smart phone, and / or a smart watch), an agricultural device (e.g., tractor and / or harvester), and / or any other type of device that may be configured to process data.

[0051] Computing system 104 may include one or more interconnects for communication between different components of the computing device. Computing system 104 may also include a network interface via which computing system 104 may exchange one or more signals including information and / or data with other computing systems and / or computing devices. For example, computing system 104 may send information and / or data (e.g., the crop cultivation instructions 124) to multi-hybrid planter 108 via the network 102. By way of further example, computing system 104 may receive information and / or data, via the network 102, from yield monitor device 106 and may acknowledge that the information and / or data was received. Further, computing system 104 may include one or more input devices (e.g., a keyboard, mouse, touch screen, stylus, and / or microphone) and / or one or more output devices (e.g., a display device and / or audio output devices including loudspeakers).

[0052] Computing system 104 may include one or more computing devices. Further, as seen in FIG. 1, computing system 104 may include one or more processors 112 and a memory 114. The one or more processors 112 may include any combination of processing devices (e.g., one or more computer processing units (CPUs), one or more graphics processing units (GPUs), one or more processor cores, one or more microprocessors, one or more application specific integrated circuits (ASICs), one or more field programmable gate arrays (FPGAs), and / or one or more controllers). By way of example, computing system 104 may comprise the one or more processors 112 and memory 114 that may store instructions that, when executed by the one or more processors 112, causes computing system 104 to perform operations which may include the operations described herein. The one or more processors 112 may execute instructions including instructions stored in the memory 114. Further, the one or more processors 112 may be arranged in various configurations including any combination of one or more serial processors and / or one or more parallel processors.

[0053] The memory 114 may include one or more computer-readable media (e.g., non- transitory computer-readable media) and may be configured to store data and / or instructions including raw raster data 120 and / or cleaned raster data 122. Further, the memory 114 may include one or more memory devices including random access memory (RAM), read only memory (ROM), programmable read only memory (PROM), erasable programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), solid state drives (SSDs), hard disk drives (HDDs), and / or hybrid memory devices that use a combination of different types of memory technologies.

[0054] Raw raster data 120 may include information associated with crop yield at a geographic location. For example, the raw raster data 120 may indicate the optimal planting density to maximize crop yield at a plurality of regions (e.g., each region having an area of one square meter) of a geographic location. Further, the raw raster data 120 may indicate information about the crop or hybrid at each geographic location. For example, the raw raster data 120 may comprise a two dimensional grid with horizontal and vertical coordinates (a set of horizontal, vertical (x, y) coordinates) corresponding to each of the regions of a geographic location. Further, each point in the raw raster data may be associated with a corresponding value that indicates a crop yield for the region of that geographic location. Each set of coordinates in the raw raster data 120 may be associated with a value that may be used to represent altitude. For example, the longitude, latitude, and altitude may be indicated by corresponding x, y, and z coordinates. Further, raw raster data 120 may be used in the generation of crop cultivation instructions 124 as described herein. Raw raster data 120 may be used in a crop breeding program to select regions of a geographic location in which crop yield is higher. Raw raster data may comprise noise that may obscure the representation of the actual geographic region on which the raw raster data is based. Noise in the raw raster data 120 may result in the misrepresentation of the actual state of the geographic location on which the raw raster data 120 is based. For example, noise in the raw raster data 120 may indicate high yield where the yield is low or low yield where the yield is high. After using the techniques described herein, the raw raster data 120 may be processed and cleaned raster data 122 may be generated. Raw raster data 120 may include the information and / or features of raw yield raster data 202 described herein with respect to FIG. 2.

[0055] Cleaned raster data 122 may include information associated with crop yield at a geographic location. The cleaned raster data 122 may be based on raw raster data 120 and mayindicate the density of crop yield for a crop at a plurality of regions (e.g., each region having an area of one square meter) of a geographic location. The cleaned raster data 122 may comprise similar information to the raw raster data 120 and may for example represent a two- dimensional or three-dimensional representation of geographic location. Further, the cleaned raster data 122 may indicate information about the crop or hybrid at each geographic location. Further, each point in the raw raster data may be associated with a corresponding value that indicates a crop yield for the region of that geographic location. Cleaned raster data 122 may be used in a crop breeding program to select regions of a geographic location in which crop yield is higher. The cleaned raster data 122 may comprise less noise than the cleaned raster data 122 and may more accurately reflect the actual conditions (e.g., yield) of the geographic location on which it is based. For example, cleaned raster data 122 may more accurately indicate high yield regions and low yield regions of a geographic location. Cleaned raster data 122 may include the information and / or features of cleaned yield raster data 204 described herein with respect to FIG. 2.

[0056] The methods described herein are not limited to cleaning harvest monitor yield data, but also yield monitor components such as grain moisture, as well as geospatially collected NIR or other data directed to grain quality analysis, such as grain oil percentage, grain protein percentage, fatty acid components of grain oil, amino acid components of grain protein, grain starch characteristics and the like. Moreover, the method can be utilized for other geospatial data such as soil type, organic matter, soil moisture, planting population, harvest population, drainage and fertility. For example, the statistical methods described herein may be applied to improve geospatial field moisture data. Field moisture data, and its components, can be expressed as a geographical raster, and so such data, like data obtained from smartfirmers or soil conductivity sensors, can be cleaned in the same manner described for the harvest monitor yield data described herein. More accurate field moisture data will translate into better management and environmental practices that involve assessing moisture levels from precipitation, evapotranspiration, irrigation, soil type, water flow, and drainage.

[0057] Further, since variations in field moisture data also result in variations in grain moisture, harvest monitors almost always take grain moisture into account when computing yield, but the methods used by harvest monitors to accomplish this can also be inaccurate or imprecise.

[0058] When geospatially cleaned moisture data obtained from impact based flow sensors were correlated with the geospatially cleaned yield data and ground truth measurements, it was surprisingly found that calibration errors of up to 20% were caused by corn grain moisture differences as small as 7% at harvest, where grain moisture levels exceeded 20% (wet basis). Similarly, in test comparisons of impact flow sensor output across a range of grain bulk densities, impact based flow sensors have been shown to experience significant inaccuracy, with errors of 5% or more experienced across a fairly modest and common range of grain bulk densities. By utilizing this cleaning method it is possible to remove non representative moisture measurements from the distribution, allowing for corrections that could reduce or remove the error in sensed yield. As sensor systems improve and more convenient means of measuring grain bulk density are developed and adopted, a similar reduction in errors due to grain bulk density could be achieved. This could allow for more accurate identification and selection of crops that dry down more quickly, use water more efficiently, produce desirable test-weights, or exhibit any of a host of grain moisture related traits. Accordingly, although not limited to this use, this method has particular benefit when used with data obtained from yield monitors with impact based flow sensors, in part because of the sensitivity of these monitors to changing moisture and grain bulk density conditions.

[0059] Likewise, nitrogen amounts, type and plant availability also vary geospatially. Accordingly, the methods described herein also can be used to correlate nitrogen practices with optimal yield. This allows for improved recommendations for nitrogen application, which recommendations can be provided as instructions to a nitrogen applicator to control the amount or type of nitrogen application, including the use of additives such as nitrogen stabilizers, which recommendations can even be tailored to a specific geographic area (such as a field or subfield management zone). From a plant breeding perspective, varieties can be selected based on their nitrogen use efficiency for a specific geographic area in which certain environmental and / or management conditions exist or are expected.

[0060] Crop cultivation instructions 124 may include information that indicate one or more crop cultivation actions that may be performed by a crop cultivation device. Further, crop cultivation instructions 124 may be used to develop a genetic variant of a crop, select certain crop varieties, evaluate traits of crop varieties, and / or distribute and multiply seeds of a crop variety. For example, crop cultivation instructions 124 may be used to direct a hybrid or crop planter device to locations in which to plant seeds based on indications of yield in cleanedraster data 122. Further, the crop cultivation instructions 124 may be used by devices at a location represented in cleaned raster data 122 to cultivate crops indicated in the crop cultivation instructions.

[0061] The yield monitor device 106 may include a mechanized and / or electronic device that may be configured to acquire and / or analyze samples of crops. In some embodiments, the yield monitor device 106 may determine phenotypic traits of crops using various sensors including image sensors (e.g., image sensors that may be used to determine the height and / or yield of a crop) and / or weight sensors (e.g., scales that may be used to determine the weight of a crop). The yield monitor device 106 may generate an indication of crop yield at a plurality of regions of a geographic location. The yield monitor device 106 may generate raw raster data 120 based on the crop samples acquired and / or analyzed by the yield monitor device 106. For example, the yield monitor device 106 may monitor and / or scan a plurality of regions of geographic location and generate raw raster data 120 comprising a two-dimensional grid with horizontal and vertical coordinates (x,y coordinates) corresponding to regions of the geographic location. The yield monitor device 106 may include any of the features and / or components of computing system 104. For example, a yield monitor device 106 may include one or more processors, a memory, one or more input devices, and / or one or more output devices. Further, a yield monitor device 106 may have different or similar configurations and / or architectures to that of computing system 104.

[0062] Multi -hybrid planter 108 may include a mechanized and / or electronic device that may be configured to perform operations comprising planting seeds of a crop. In some embodiments, multi-hybrid planter 108 may plant seeds of one or more crops comprising rice, wheat, potatoes, barley, corn, soybeans, and / or cotton. The multi -hybrid planter 108 may be configured to receive one or more signals which may comprise crop cultivation instructions based on the crop cultivation instructions 124. Further, multi-hybrid planter 108 may be configured to plant one or more seeds of a crop indicated in the crop cultivation instructions 124. For example, multi-hybrid planter 108 may plant wheat seeds in a location, at a quantity, and at a time indicated in the crop cultivation instructions. The multi -hybrid planter 108 may include any of the features and / or components of computing system 104. For example, a multihybrid planter 108 may include one or more processors, a memory, one or more input devices, and / or one or more output devices. Further, a multi -hybrid planter 108 may have different or similar configurations and / or architectures to that of computing system 104.

[0063] Irrigation system 110 may include a mechanized device, electronic device, and / or structures that may be configured to irrigate crops. In some embodiments, the irrigation system 110 may irrigate crops using water and / or other substances that may be used to promote or inhibit the growth of crops. The irrigation system 110 may perform surface irrigation (e.g., furrow irrigation) that irrigates crops through use of water channels, localized irrigation in which crops may be irrigated using pipes, drip irrigation in which hoses may be configured to drip water onto crops, and / or sprinkler irrigation in which crops are irrigated by being sprayed with water. The crop cultivation instructions 124 may be used to determine the configuration of the irrigation system 110. For example, the crop cultivation instructions may indicate different amounts of water and / or fertilizer to provide to different regions of a geographic location. Irrigation system 110 may include any of the features and / or components of computing system 104. For example, an irrigation system 110 may include one or more processors, a memory, one or more input devices, and / or one or more output devices. Further, an irrigation system 110 may have different or similar configurations and / or architectures to that of computing system 104.

[0064] One or more aspects described herein may be embodied in computer-usable, computer-readable data, and / or computer-executable instructions, which may be stored as data and / or instructions in one or more memory devices and / or executed by one or more computing devices and / or other devices described herein. Data and / or instructions may include software applications and / or computer programs that may be used to perform the operations described herein (e.g., generating the cleaned raster data 122) when executed by one or more processors in a computing device or other device. The data and / or instructions may be written in source code that is compiled for execution by a computing device. The computer executable instructions may be stored on a computer readable medium (e.g., a non -transitory computer- readable medium) such as a hard disk, solid state drive, optical disk, removable storage media, solid state memory, and / or RAM. The functionality of the computing applications described herein may be combined or distributed in various embodiments. Further, the functionality of the computing applications described herein may be partly or wholly embodied in firmware or hardware equivalents including integrated circuits and / or field programmable gate arrays (FPGA).

[0065] FIG. 2 shows an example of raw yield data, cleaned yield data, and yield difference data according to one or more aspects of the disclosure. The raw yield raster data, cleaned yieldraster data, and yield difference map may be generated by the systems and devices described herein including the system 100 shown in FIG. 1.

[0066] FIG. 2 raw yield raster data 202 may comprise information that is used to generate an image that corresponds to a representation of a geographic location. The raw yield raster data 202 may be based on an image captured by one or more sensors that are used to detect the state of a geographic location. For example, raw yield raster data 202 may comprise satellite imagery and / or aerial imagery of a crop field that is captured by one or more cameras. As shown in FIG.2, the region 214 comprises noise that may obscure the actual state of the geographic location on which raw yield raster data 202 is based. The noise in the region 214 may comprise one or more image artifacts.

[0067] The cleaned yield raster data 204 and 206 for hybrids 1 and 2 (or varieties 1 and 2 for non-hybrid crops) may be based on the areas of raw yield raster data 202 where hybrid 1 and 2 were planted and may comprise an image of the same geographic locations indicated in the raw yield raster data 202. The cleaned yield raster data 204 and 206 may be the result of one or more operations to denoise the raw yield raster data 202 by detecting and removing spatial outliers and / or statistical outliers in the raw yield raster data 202 for each hybrid. As shown in FIG.2, the region 216 corresponds to the region 214 in the raw yield raster data 202 and has been denoised, which may allow the actual state of the underlying geographic location to be more accurately represented.

[0068] The kriged yield data 208 and 210 for hybrids 1 and 2 may be based upon the cleaned yield raster data 204 and 206, along with the cleaned variogram indicated by the spatial cleaning procedure for each hybrid. Locations where spatial or statistical outliers were removed, as well as locations where the alternate hybrid was planted, may be replaced with kriged yield estimates for each hybrid across all geographical locations represented in raw yield raster data 202. The yield difference data 212 may indicate differences between the kriged yield data 208 and 210 for hybrids 1 and 2, such as specific geographic regions where one hybrid outperformed the other.

[0069] FIG. 3 shows a timing diagram that indicates a data flow of signals associated with removal of noise from geographical raster datasets according to one or more aspects of the disclosure. Computing devices and / or systems associated with data flow 300 may include any of the features and / or components of computing system 104, yield monitor device 106, multi-hybrid planter 108, and / or irrigation system 110, which are shown in FIG. 1. Further, the operations performed as part of the data flow 300 may be performed by the computing systems and / or computing devices that are described herein.

[0070] Yield monitor device 106 may detect the state of crops and generate raw raster data (e.g., the raw raster data 120 described with respect to FIG. 1) based on the detected states (e.g., yield based on in-season remote sensing) of the crops. Yield monitor device 106 may then send a signal 312 that includes the raw raster data to computing system 104.

[0071] Computing system 104 may be configured to receive signals comprising the signal 312, the signal 316, and / or the signal 320. In some embodiments, computing system 104 may be configured to send signals comprising a request for data (e.g., raw raster data) to yield monitor device 106. In response to the request for data from computing system 104, yield monitor device 106 may send raw raster data to computing system 104. The computing system 104 may process the raw raster data received from the yield monitor device 106 and may generate cleaned raster data (e.g., the cleaned raster data 122 described with respect to FIG. 1). Further, the computing system 104 may use the cleaned raster data 122 to generate crop cultivation instructions that may be used in a crop breeding program that uses the cleaned raster data 122 in the selection of crop varieties based on the yield of the crop variety. The cultivation instructions may then be sent to one or more cultivation devices that may be used to cultivate the selected crop varieties.

[0072] For example, the computing system 104 may send the signal 314 to multi-hybrid planter 108. The signal 314 may comprise crop cultivation instructions indicating seeds of a crop variety to plant at specified regions of a geographic location. For example, the multihybrid planter 108 may receive the signal 314 and plant seeds of a specified crop variety at regions of a location indicated in the crop cultivation instructions. Based on receiving the signal 314, multi-hybrid planter 108 may send signal 318 (e.g., a signal confirming receipt of the crop cultivation instructions indicated in the signal 314) to computing system 104. Multi-hybrid planter 108 may use the crop cultivation instructions included in the signal 316 to plant seeds of the crop variety specified in the crop cultivation instructions.

[0073] Further, the computing system 104 may send the signal 318 to irrigation system 110. The signal 318 may comprise crop cultivation instructions indicating regions of a geographic area to which specified amounts of water, fertilizer, and / or pesticide may bedistributed. For example, the irrigation system 110 may receive the signal 318 and distribute water to specified regions of a geographic location indicated in the crop cultivation instructions. Based on receiving the signal 318, irrigation system 110 may send signal 318 (e.g., a signal confirming receipt of the crop cultivation instructions indicated in the signal 318) to computing system 104.

[0074] FIG. 4 shows an example flow chart for raster cleaning and error estimation of spatial data according to one or more aspects of the disclosure. One or more aspects of the disclosure may be implemented by the devices described herein (e.g., the system 100 shown in FIG. 1). One or more of the steps described with respect to FIG. 4 may be omitted, performed in a different order, and / or modified. Further, one or more additional steps (e.g., the steps described with respect to FIGS. 5-8) may be added to the steps described with respect to FIG. 4.

[0075] At step 402, a variogram may be generated. For example, the computing system 104 may generate a variogram based on raw raster data (e.g., the raw raster data 120 described with respect to FIG. 1). The variogram may indicate the variation in values (e.g., yield values) between two regions of a geographic location based on the distance between the two regions. The variogram may indicate that the yield at locations that are close to one another (e.g., adjacent) may vary less than the yield at locations that are further apart (e.g., regions separated by hundreds of meters).

[0076] At step 404, statistical outliers may be detected. For example, the computing system 104 may analyze raster data to detect statistical outliers that exceed certain thresholds. For example, values (e.g., yield values) outside the range of 5 to 500 bushels per acre may be determined to be statistical outliers. In some embodiments, a two-step process is used, for example removal of an absolute range (e.g. yield values outside a specified range, such as 5 to 500 bushels per acre), followed by the removal of remaining points more than a specified statistical distance from a mean, median or other desired point at step 406 (e.g. three standard deviations from the resulting mean). The number of standard deviations can be optimized to any value (e.g. 2, 3, 4 standard deviations from the mean) as needed for any data set. This two- step process keeps extreme outliers from skewing the standard deviation computation.

[0077] At step 408, a recursive stack may be created by providing all of the spatial data to the spatial filter function. As the stack is not empty at step 410, the top dataset on the stack(the original dataset) may be removed from the top of the stack at step 412 and the number of observations provided to the stack may be evaluated against a threshold value at step 414.

[0078] If the number of observations is less than the threshold value, the spatial filtering function may be applied to the current dataset, with spatial outliers identified and marked, and with filtering done, the resulting dataset may be moved to final storage. If at step 414 the number of observations exceeds the threshold, at step 416 the data may be divided into four quadrants based upon their median x and y coordinates, and those four datasets may be sequentially placed on the stack, awaiting their turn to be spatially filtered.

[0079] This process may continue recursively until each dataset meets the threshold size and step 418 has been performed on each dataset in the stack. The observation threshold may be based on the available computational resources to process the spatial outlier detection operations. Greater availability of computational resources may be correlated with larger threshold values.

[0080] At step 420, a clean variogram may be generated. The clean variogram may be generated based on clean raster data (e.g., the clean raster data described with respect to FIG.1) in which spatial outliers and / or statistical outliers that were detected have been removed. As a result, the clean variogram may more accurately reflect the true spatial weighting pattern of yield in the geographic location on which the variogram is based.

[0081] At step 422, raster locations where errors were removed by either the statistical or spatial filtering methods can be replaced using kriging based upon the spatial weighting specified by the clean variogram. Further, errors associated with the kriging process at those locations may be estimated.

[0082] FIG. 5 shows an example flow chart for spatial outlier detection according to one or more aspects of the disclosure. One or more aspects of the disclosure may be implemented by the devices described herein (e.g., the system 100 shown in FIG. 1). One or more of the steps described with respect to FIG. 5 may be omitted, performed in a different order, and / or modified. Further, one or more additional steps (e.g., the steps described with respect to FIGS. 4 and / or 6-8) may be added to the steps described with respect to FIG. 5.

[0083] At step 502, there may be a determination of whether all processing of all locations of a geographic area indicated in raw raster data (e.g., the raw raster data 120) have beencompleted. Based on determining that all locations of a geographic area have been completed, the method may end. Based on determining that all locations have not been completed, steps 504-522 may be performed until processing of all locations has been completed. Performing steps 504-522 in a recursive manner may result in a reduction in complexity from O(N2) to O(N log N), thereby reducing the processing resources and / or time required to perform the steps 504-522.

[0084] At step 504, The location may be incremented to a location that has not already been analyzed. For example, if a starting location z has been processed, the next location would be z+1, followed by z+2, until all locations have been processed.

[0085] At step 506, K neighbors of location z may be identified. For example, the K neighbors of location z may comprise locations that are adjacent to location z.

[0086] At step 508, distance-based weights may be applied. For example, neighbors that are a greater distance from the location z may be weighted less heavily than neighbors that are closer to the location z.

[0087] At step 510, weights may be row standardized. For example, the weights determined in step 508 may be standardized by dividing each weight by the sum of the weights of neighboring elements.

[0088] At step 512, z-scores for neighbors may be determined. For example, z-scores may be determined for neighbors that are adjacent to location z. The z-scores for the neighbors may indicate each neighbors value relative to the mean of the weights of the other neighbors.

[0089] At step 514, Moran’s I values may be determined to correspond to the product of a determined spatial lag and the z-scores for the location z. The Moran’s I values may indicate spatial autocorrelation between location z and / or the neighbors of location z.

[0090] At step 516, a p-value for Moran’s I values greater than zero may be determined. The p-values may indicate a probability of obtaining results that are at least equal to the results that were observed.

[0091] At step 518, The location may be deleted if Moran’s I is less than 0 and the p-score is greater than 0.05. As a result a spatial outlier may be removed from the raw raster data, resulting in cleaned raster data.

[0092] At step 520, Cook’s distance may be determined for the location i. The Cook’s distance may indicate the leverage and / or residual of each location when the location is removed.

[0093] At step 522, a location may be deleted if Cook’s distance for location z is greater than 4 / N (e.g., N locations) and the lag is less than 0. After performing step 522, the method may return to the step 502.

[0094] FIG. 6 shows an example flow chart for error estimation according to one or more aspects of the disclosure. One or more aspects of the disclosure may be implemented by the devices described herein (e.g., the system 100 shown in FIG. 1). One or more of the steps described with respect to FIG. 6 may be omitted, performed in a different order, and / or modified. Further, one or more additional steps (e.g., the steps described with respect to FIGS. 4, 5, 7, and / or 8) may be added to the steps described with respect to FIG. 6.

[0095] At step 602, a covariogram may be generated. For example, the computing system 104 may generate a covariogram based on a variogram (e.g., the variogram of step 402 described with respect to FIG. 4). The covariogram may indicate the covariance in values among regions indicated in raster data.

[0096] At step 604, a minimum window size may be determined. The minimum window size may be used to reduce the search space of raster data to facilitate kriging (e.g., the kriging performed in step 412). The minimum window size may indicate the number of neighbors that may be included within a window. A greater minimum window size may require a greater amount of computing resources and / or time to process.

[0097] At step 606, a window may be extracted for N neighbors. For example, a window based on the minimum size may be used to extract N neighbors.

[0098] At step 608, there may be a determination of whether N is greater than a minimum N threshold value. Based on N not being greater than a minimum N threshold value, step 610 may be performed. Based on N being greater than a minimum N threshold value, step 612 may be performed.

[0099] At step 610, the search window size may be increased. The search window size may be increased incrementally such that the search window gradually increases until the minimum N threshold value is reached.

[0100] At step 612, a yield estimate (e.g., or any other z value) at location x, y (e.g., a location at a set of coordinates x, y) may be estimated. For example, the yield estimate (z) for location x, y may indicate the location’s value relative to a weighted mean of the neighboring locations.

[0101] At step 614, an error based on the geometry of neighboring locations and a local error may be determined for location x, y. This error may indicate the degree to which the locations of the neighbors are near and / or evenly surrounding the target location. At step 616, a localized error of the neighborhood at location x, y may be determined.

[0102] At step 618, there may be a determination of whether location x, y (e.g., the current location being processed) is the last remaining location that has not been processed. Based on the current location not being the last location, step 620 may be performed. Based on the current location being the last location, the method may end.

[0103] FIG. 7 shows an example flow chart for raster cleaning according to one or more aspects of the disclosure. One or more aspects of the disclosure may be implemented by the devices described herein (e.g., the system 100 shown in FIG. 1). One or more of the steps described with respect to FIG. 7 may be omitted, performed in a different order, and / or modified. Further, one or more additional steps (e.g., the steps described with respect to FIGS. 4-6, and / or 8) may be added to the steps described with respect to FIG. 7.

[0104] At step 702, raw raster data may be received and / or acquired (e.g., computing system 104 may request raw raster data from a computing device that generates and / or stores the raw raster data). The raw raster data may comprise a plurality of points corresponding to a plurality of portions of a geographic area. The plurality of points may be associated with a plurality of yield values in the plurality of portions of the geographic area. For example, computing system 104 may receive raw raster data from yield monitor device 106.

[0105] At step 704, based on detection of a plurality of points that are statistical outliers, step 706 may be performed and the plurality of points that meet one or more statistical outlier criteria may be removed from the raw raster data. For example, computing system 104 mayanalyze the raw raster data and detect statistical outliers in the raw raster data. Based on a plurality of points not meeting the one or more statistical outlier criteria, step 702 may be performed.

[0106] At step 706, the plurality of points that meet the one or more statistical outlier criteria may be removed. For example, the computing system 104 may remove statistical outliers comprising the plurality of points that exceed a spatial outlier threshold.

[0107] At step 708, based on application of one or more statistical outlier criteria to the raw raster data, a plurality of regions may be determined based on the raw raster data from which the statistical outliers have been removed. Each of the plurality of regions may comprise a different set of the plurality of points and / or each of the plurality of regions comprises fewer points than a threshold number of the plurality of points. For example, computing system 104 may apply one or more statistical outlier thresholds to the raw raster data.

[0108] At step 710, based on application of one or more spatial outlier detection techniques to each of the plurality of regions, spatial outliers in the plurality of regions may be detected (e.g., recursively detected). The one or more spatial outlier detection techniques may comprise one or more algorithms and / or analysis techniques that may be used to detect spatial outliers. For example, the one or more spatial outlier detection techniques may comprise spatial clustering analysis of the plurality of points in each of the plurality of regions. For example, computing system 104 may analyze the raw raster data and use spatial clustering analysis to detect spatial outliers in the raw raster data.

[0109] At step 712, a raster variogram may be generated. The raster variogram may be based on the plurality of regions from which the spatial outliers have been removed. For example, computing system 104 may generate a raster variogram based on the raw raster data from which spatial outliers were removed.

[0110] At step 714, based on the raster variogram and / or application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data may be generated. For example, computing system 104 may generate clean raster data based on the application of kriging techniques to the raw raster data from which the statistical outliers and / or spatial outliers were removed. The one or more kriging techniques may result in a best linear unbiased prediction (BLUP) for the plurality of regions. Sometimes kriging is referred to as a Gaussian process regression; it is a method of interpolation based onGaussian process governed by prior covariances, and under suitable assumptions of the prior, kriging gives the best linear unbiased prediction (BLUP) at unsampled locations.

[0111] At step 716, based on the clean raster data, the crop variety that is highest yielding for the geographic area may be selected. For example, the computing system may analyze the clean raster data and compart regions of the geographic area to determine the crop variety that is the highest yielding.

[0112] At step 718, an indication for a plant breeding program to advance development of the crop variety that is the highest yielding may be generated. For example, an indication identifying the highest yielding crop variety may be generated for use in a plant breeding program. The plant breeding program may perform population development, breeding crosses, hybrid testing, and / or variety testing on the crop variety that is the highest yielding. In some embodiments, a message comprising the indication of the crop variety that is the highest yielding may be sent to a computing device of a crop breeding program.

[0113] At step 720, based on the clean raster data and the crop variety that is the highest yielding, crop cultivation instructions to cultivate the crop variety that is the highest yielding may be generated. For example, the computing system 104 may generate crop cultivation instructions to plant seeds of the highest yielding crop at a plurality of locations indicated in the cleaned raster data.

[0114] FIG. 8 shows an example flow chart for use of cultivation instructions according to one or more aspects of the disclosure. One or more aspects of the disclosure may be implemented by the devices described herein (e.g., the system 100 described with respect to FIG. 1). One or more of the steps described with respect to FIG. 8 may be omitted, performed in a different order, and / or modified. Further, one or more additional steps (e.g., the steps described with respect to FIGS. 4-7) may be added to the steps described with respect to FIG. 8.

[0115] At step 802, crop cultivation instructions may be received. The crop cultivation instructions may be used to operate devices used in a crop breeding program. For example, the crop cultivation instructions may be sent to a hybrid-planter that may plant one or more quantities of a crop variety based on the crop cultivation instructions. A planting device (e.g., the multi-hybrid planter 108 described with respect to FIG. 1) may comprise a wireless receiver that is configured to receive signals comprising crop cultivation instructions.

[0116] For example, the multi-hybrid planter 108 may receive crop cultivation instructions indicating specific amounts of a particular genotype of com to plant at a specific location at a specific time and / or date. In some embodiments, the crop cultivation instructions may be encrypted and the multi -hybrid planter 108 may decrypt the crop cultivation instructions and then perform operations indicated in the crop cultivation instructions after the sender of the crop cultivation instructions has been authenticated.

[0117] At step 804, one or more quantities of seeds of the one or more combinations of the plurality of crop varieties may be planted at one of the plurality of locations. The quantities, types of seeds, and locations to plant the seeds may be indicated in the crop cultivation instructions.

[0118] Initial cleaning of nearly 100,000 rasterized raw datasets using the method described herein resulted in significant and surprising levels of statistical improvements.

[0119] With regard to yield data sets, yield variance was reduced by a median 54%, with a few extreme outliers driving the mean yield variance reduction above 90%. Cleaning increased average yield by a median of 3.8 bushels per acre, while extreme outliers pushed the mean increase to 9.8 bushels per acre. Spatial signal to noise ratios, when computing using the nugget and sill of the raw and clean variograms improved a median 20x with a mean of 24x.

[0120] In particular, yield data with grain moisture differences were significantly improved with the method described, as yield data can be particularly susceptible to errors that have little to no means of identification or removal with standard methods. For example, grain moisture differences can cause 15-20% or larger errors in yield calibrations when moisture levels significantly exceed 20% and include large (5-7%) deviations in moisture level. The methods developed here allow for cleaning moisture data (removing spurious moisture errors) and provides detection of areas where combine calibration errors are likely to be excessive. This allows for the potential correction or removal of data where wild swings at high moisture levels are likely to be problematic for the underlying yield estimates.

[0121] With regard to ‘as planted’ seed planting density data, data obtained from the seed planter was cleaned using the method described, and significant and surprising effects were also found when the method was applied to this data set. The median population variance reduction was 90%, with the mean reduction approaching 97% due to extreme outliers. The cleaning reduced the median planting population by 169 seeds per acre, with the mean reducedby 449 seeds per acre. Spatial signal to noise ratios for population increased a median of lOx, with a mean of l lx. Similar results could be expected with cleaned ‘as applied’ data sets of crop input applications such as fertilizer, nitrogen stabilizers, herbicides, biologicals, fungicides or insecticides. Likewise, the cleaned method as applied to as planted inbred seed data would be useful information for optimizing hybrid seed production practices and predicting potential hybrid seed yield.

[0122] The method described herein is also particularly useful in providing a high quality clean data set for artificial intelligence models, whether machine learning models or deep learning, and whether supervised or unsupervised.

[0123] For example, hybrid or variety seed selection models require a huge amount of training data to be able to predict hybrid performance across a wide range of soil conditions and weather environments. Traditionally, the means of acquiring this data has been the standard strip trial, where a farmer locates a relatively “fixed” soil environment on a field and applies strips of various test hybrids or varieties. The data collected for each of these strip trial locations, along with soil, weather and satellite data, are combined into training sets and machine learning techniques are used to train model inputs to the strip trial measured outputs. This process can be updated each season as new strip trial data becomes available. However, this approach tends to underestimate the amount of true environmental variation that exists in farmers’ fields, as the strip trial sites tend to be biased towards better soil environments, for a variety of reasons.

[0124] On farm split planter trials (where two hybrids or varieties are planted in adjacent narrow strips across a field, and then harvested as individual products) provide an opportunity to compare hybrids or varieties across thousands of additional soil, site and management conditions. This greatly increases the ability of artificial intelligence algorithms to encode and express the full variability inherent across farms, states and regions. The use of the data cleaned according to the method described herein will produce a significantly clearer and more accurate spatial model across landscapes than a model that uses yield data that has not been cleaned according to this method, and significantly enhances the value of on farm split planter trial data.

[0125] Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject matter defined in theappended claims is not necessarily limited to the specific features or acts described above. While embodiments described primarily envision use of the method with row crops such as corn, soybean, wheat, rice, canola, potatoes, sorghum and the like, the specific features and acts described above are disclosed as example forms of implementing the claims, and use of the method with other vegetables, ornamentals or other crops is also envisioned, whether annual, perennial, shrub, bush or tree. The steps of the methods described herein are described as being performed in a particular order for the purposes of discussion. A person having ordinary skill in the art will understand that the steps of any methods discussed herein may be performed in any order and that any of the steps may be omitted, combined, and / or expanded without deviating from the scope of the present disclosure. Furthermore, the methods described herein may be performed using any manner of device, system, and / or apparatus including the computing devices, computing systems, and / or computing apparatuses that are described herein.

Claims

What is claimed is:

1. A system for objectively selecting a highest yielding crop variety in a geographic area comprising: a mechanized yield monitor configured to: determine a plurality of yield values of one or more crops within a plurality of portions of the geographic area; generate raw raster data comprising a plurality of points corresponding to the plurality of portions of the geographic area, wherein the plurality of points are associated with the plurality of yield values in the plurality of portions of the geographic area; a computing device configured to: receive the raw raster data; remove, from the raw raster data, the plurality of points that meet one or more statistical outlier criteria; determine, based on application of one or more spatial outlier criteria to the raw raster data from which statistical outliers have been removed, a threshold point density of the plurality of points to use in detection of spatial outliers; determine a plurality of regions based on the raw raster data from which the statistical outliers have been removed, wherein each of the plurality of regions comprises a different set of the plurality of points, and wherein each of the plurality of regions has a lower point density than the threshold point density; recursively remove, from the plurality of regions, based on application of one or more spatial outlier detection techniques to each of the plurality of regions, the plurality of points that meet one or more spatial outlier criteria, wherein the one or more spatial outlier detection techniques comprise spatial clustering analysis of the plurality of points in each of the plurality of regions; generate a raster variogram based on the plurality of regions from which the spatial outliers have been removed; generate, based on the raster variogram and application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data; select, based on the clean raster data, the crop variety that is highest yielding for the geographic area; andgenerate an indication for a plant breeding program to advance development of the crop variety that is the highest yielding.

2. The system of claim 1, wherein the system is further configured to: recommend the selected highest yielding crop variety for planting in a location with environmental characteristics that match the geographic area in which the selected highest yielding crop variety was evaluated.

3. The system of claim 2, wherein the recommending comprises recommending a multihybrid planter to plant seeds using crop cultivation instructions.

4. The system of claim 1, wherein the system is configured to: generate an image of the geographic area based on the clean raster data, wherein the image comprises a plurality of indications based on the plurality of yield values associated with the plurality of portions of the geographic area.

5. The system of claim 4, wherein the one or more crops comprise at least two hybrid crop varieties, and wherein the plurality of indications indicate which of the at least two hybrid crop varieties is planted in each of the plurality of portions of the geographic area.

6. The system of claim 1, wherein the mechanized yield monitor determines the plurality of yield values using an impact flow sensor.

7. The system of claim 1, wherein the plurality of yield values are obtained from a split planter trial.

8. A method of selecting and planting a crop based on raster data, the method comprising: acquiring, by a computing device comprising one or more processors, raw raster data comprising a plurality of points corresponding to a plurality of portions of a geographic area, wherein the plurality of points are associated with a plurality of yield values in the plurality of portions of the geographic area;based on detection of the plurality of points that are statistical outliers, removing, by the computing device, from the raw raster data, the plurality of points that meet one or more statistical outlier criteria; determining, by the computing device, based on application of one or more statistical outlier criteria to the raw raster data, a plurality of regions based on the raw raster data from which the statistical outliers have been removed, wherein each of the plurality of regions comprises a different set of the plurality of points, and wherein each of the plurality of regions comprises fewer points than a threshold number of the plurality of points; recursively detecting, by the computing device, based on application of one or more spatial outlier detection techniques to each of the plurality of regions, spatial outliers in the plurality of regions, wherein the one or more spatial outlier detection techniques comprise spatial clustering analysis of the plurality of points in each of the plurality of regions; generating, by the computing device, a raster variogram based on the plurality of regions from which the spatial outliers have been removed; generating, by the computing device, based on the raster variogram and application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data; selecting, by the computing device, based on the clean raster data, a crop variety that is highest yielding for the geographic area; generating, by the computing device, an indication for a plant breeding program to advance development of the crop variety that is the highest yielding; and generating, by the computing device, based on the clean raster data, crop cultivation instructions to cultivate the crop variety that is the highest yielding.

9. The method of claim 8, further comprising: correlating, by the computing device, the crop cultivation instructions with historical crop growth factors for the geographic area, the historical crop growth factors for the geographic area comprising daily precipitation, daily evapotranspiration, irrigation practices, or soil type; and determining, by the computing device, based on the correlation between the crop cultivation instructions and the historical crop growth factors, the one or more crops to cultivate in each of the plurality of portions of the geographic area.

10. The method of claim 8, further comprising:providing the crop cultivation instructions to one or more crop cultivation devices configured to cultivate the one or more crops based on the crop cultivation instructions.

11. The method of claim 10, wherein the one or more crop cultivation devices comprise a multi-hybrid planter or an irrigation device.

12. The method of claim 8, wherein the crop cultivation instructions comprise plant density prescriptions.

13. The method of claim 8, wherein the determining the plurality of amounts of the one or more crops to cultivate is positively correlated with the plurality of yield values in the plurality of portions of the geographic area.

14. The method of claim 8, wherein the plurality of yield values were obtained from an impact flow sensor.

15. The method of claim 8, wherein the plurality of yield values were obtained from a split planter trial.

16. The method of claim 8, wherein the one or more statistical outlier criteria comprise a minimum yield value or a maximum yield value.

17. The method of claim 16, wherein the minimum yield value or the maximum yield value are based on a standard deviation of the plurality of yield values.

18. The method of claim 8, wherein the one or more spatial outlier detection techniques are performed serially on one of the plurality of regions at a time.

19. The method of claim 8, wherein the one or more spatial outlier detection techniques are performed in parallel on more than one of the plurality of regions at a time.

20. The method of claim 8, wherein the one or more spatial outlier detection techniques comprise determining Moran’s I for the plurality of points in a plurality of nonoverlapping regions.

21. The method of claim 8, wherein the one or more spatial outlier detection techniques comprise determining a plurality of Cook’s distances based on the plurality of yield values associated with the plurality of points in the plurality of regions.

22. The method of claim 21, wherein the spatial outliers comprise the plurality of points corresponding to the plurality of Cook’s distances greater than a predetermined Cook’s distance threshold.

23. The method of claim 8, further comprising: determining an amount of computational resources that are available; and determining the threshold number of the plurality of points to analyze recursively based on the amount of computational resources available.

24. The method of claim 8, wherein the plurality of points in each of the plurality of regions are mutually exclusive.

25. A method of generating cleaned raster data, the method comprising: acquiring, by a computing device comprising one or more processors, raw raster data comprising a plurality of points corresponding to a plurality of portions of a geographic area, wherein the plurality of points are associated with a plurality of geospatial values in the plurality of portions of the geographic area; determining, by the computing device, based on removal of statistical outliers from the raw raster data, a plurality of regions comprising fewer points than a threshold number of the plurality of points; detecting, by the computing device, based on application of one or more spatial outlier detection techniques to each of the plurality of regions, spatial outliers in the plurality of regions, with a number and size of regions determined by recursive division of coordinate space to meet a maximum size threshold, wherein the one or more spatial outlier detection techniques comprise spatial clustering analysis of the plurality of points in each of the plurality of regions; generating, by the computing device, a raster variogram based on the plurality of regions from which the spatial outliers have been removed; andgenerating, by the computing device, based on the raster variogram and application of one or more kriging techniques to the plurality of regions from which the spatial outliers have been removed, clean raster data.

26. The method of claim 25, wherein the one or more spatial outlier detection techniques comprise determining Moran’s I for the plurality of points in a plurality of nonoverlapping regions.

27. The method of claim 25, wherein the one or more spatial outlier detection techniques comprise determining a plurality of Cook’s distances based on the plurality of geospatial values associated with the plurality of points in the plurality of regions.

28. The method of claim 27, wherein the spatial outliers comprise the plurality of points corresponding to the plurality of Cook’s distances greater than a predetermined Cook’s distance threshold.

29. The method of claim 25, wherein the plurality of geospatial values are yield values that were obtained from an impact flow sensor.

30. The method of claim 25, wherein the plurality of geospatial values are yield values that were obtained from a split planter trial.

31. The method of claim 25, wherein the cleaned raster data is used to select a variety to plant in a specific geographic area.

32. The method of claim 25, wherein the cleaned raster data is utilized to select agronomic practices to utilize for a selected variety in the geographic area.

33. The method of claim 25, wherein the plurality of geospatial values are used as planted seed planting density values.

34. The method of claim 25, wherein the plurality of geospatial values are used as applied values of a crop input.

Images