Automatic phenotype evaluator

Abstract

A phenotype evaluation platform for evaluating a plant phenotype is provided. The phenotype evaluation platform includes a scale, a photography platform, optical and depth sensors, and a calibrated light source. A configuration puck including a plant type identifier, a QR code, and instructions for preparing the plant, a plant identification card including a plant identifier, and a color calibration card including a color swatch of a known value are positioned on the photography platform. The plant is positioned according to the instructions, illuminated with the calibrated light source, and optical and depth images are captured. A metadata file generated based on the images includes the QR code and the plant identifier. A digital color code of the plant color is extracted, a plant depth is calculated, a plant phenotype is determined based at least on the digital color code and the depth and displayed on a computer display.

Description

AUTOMATIC PHENOTYPE EVALUATORCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 558,841 filed February 28, 2024, entitled “AUTOMATIC PHENOTYPE EVALUATOR”, all of which is hereby incorporated by reference in its entirety.TECHNICAL FIELD

[0002] Various of the disclosed embodiments concern an automatic phenotype evaluator.BACKGROUND

[0003] When growing plants, there are many decisions that must be made such as how much water, light, or nutrients to provide the plants. Other determinations such as which variety grows best, which treatment for pests works best, or which growing strategy works best must also be made. To make these determinations, it is useful to collect data about the plants by measuring specific traits or attributes of the plant such as weight, size, leaf or fruit color, growth, morphology, architecture, and composition. This process is known as plant phenotyping. This data can be used to determine whether alterations in the growing environment are needed to encourage better plant growth. It may also be useful to measure some attributes that affect consumer demand for an agricultural plant such as leaf straightness. These attributes may be measured during the growing cycle or at harvest time.

[0004] Historically, these measurements were made using a manual process such as weighing each head of lettuce in a harvest and recording the weight in a journal. Similarly, leaf color was subjectively evaluated by a person and recorded in some fashion. This manual process was tedious, slow, and prone to errors. These errors may result from human subjectivity as in leaf color, data entry errors, or difficulty in quantifying, such as in plant size. Moreover, the process did not have a means for verifying the data after the fact.

[0005] There are also some data or measurements that historically were infrequently collected such as recording the location of the plant in relation to otherplants. This information could be used to determine if some growing areas are better than others.

[0006] If phenotyping were easier and more accurate, it could be used for plant research, quality control, and yield predictions.SUMMARY

[0007] Embodiments of the invention provide a more streamlined process for gathering data about plants and crops that produces more accurate and precise data free of human subjectivity.BRIEF DESCRIPTION OF THE DRAWINGS

[0008] One or more embodiments of the present disclosure are illustrated by way of example and not limitation in the figures of the accompanying drawings, in which like references indicate similar elements.

[0009] Figure 1 is a perspective view of the herein disclosed inventive platform.

[0010] Figures 2 - 24 provide examples of configuration pucks according to the invention.

[0011] Figure 25 is a block diagram of a computer system as may be used to implement certain features of some of the embodiments.

[0012] Figure 26 is a block diagram that illustrates an example artificial intelligence (Al) system that can implement aspects of the present technology.

[0013] Figure 27 is a block diagram that illustrates an example of a computer system in which at least some operations described herein can be implemented.

[0014] Figure 28 is a block diagram that illustrates an example of a system in which at least some aspects of the present technology are implemented.

[0015] Figure 29 is a block diagram of a process in which at least some aspects of the disclosed technology are implemented.

[0016] Figure 30 shows a flowchart of a process 3000 in which aspects of the disclosed technology are implemented.DETAILED DESCRIPTION

[0017] Various example embodiments will now be described. The following description provides certain specific details for a thorough understanding and enabling description of these examples. One skilled in the relevant technology will understand, however, that some of the disclosed embodiments may be practiced without many of these details.

[0018] Likewise, one skilled in the relevant technology will also understand that some of the embodiments may include many other obvious features not described in detail herein. Additionally, some well-known structures or functions may not be shown or described in detail below to avoid unnecessarily obscuring the relevant descriptions of the various examples.

[0019] The terminology used below is to be interpreted in its broadest reasonable manner, even though it is being used in conjunction with a detailed description of certain specific examples of the embodiments. Indeed, certain terms may even be emphasized below; however, any terminology intended to be interpreted in any restricted manner will be overtly and specifically defined as such in this Detailed Description section.Automatic Phenotype Evaluator

[0020] Embodiments of the invention provide a platform that includes all the necessary equipment for phenotyping and logging plant data.

[0021] Figure 1 is a perspective view of the herein disclosed inventive platform.

[0022] In some embodiments, a wheeled cart 100 can be provided that can be wheeled around inside and down the rows of a greenhouse. The cart can contain all the necessary equipment for phenotyping and logging plant data. Alternate embodiments of the invention may use a more rugged cart with larger off-road tires capable of navigating an outdoor farm field. Embodiments of the invention may also include powered wheels coupled to an energy source to allow the wheeled cart 100 to move under its own power or the wheeled cart 100 may contain self-driving capabilities that allow the wheeled cart 100 to navigate on its own. Embodiments of the invention may also be permanently affixed in a conveyor belt system to allow it to be used on all products moving down the conveyor belt system.

[0023] To decrease data entry errors, a computer can be included to interface with all of the equipment that is used to retrieve the necessary data. The equipment may include a scale 101 , a depth sensor / camera 106, a calibrated light source 102, a barcode or QR code scanner, a keyboard 104, or set of buttons 107 (also referred to herein as button panel 107), and a display 105 that can provide a user interface. Additionally, chemical analysis can be performed as well, such as with a brix meter, also known as a refractometer, to measure the sugar content of the sample or a pH meter to measure the acidity or alkalinity of the sample. Each of these devices directly interfaces with the computer to prevent any chance for a human to cause data entry errors. Analysis of the acquired data may be conducted on the cart 100 itself or may be handled by an offline system such as a cloud infrastructure.

[0024] In embodiments of the invention, a photography platform 103 can provide a flat surface of a consistent uniform color free from any patterns or other visual marks so that a photographic mask can be easily calculated. The photography platform 103 can be placed in the field of view of the depth sensor / camera 106 so that the entire photography platform 103 is in focus. The photography platform 103 can be placed on top of the scale 101 if the scale 101 is used.

[0025] The depth sensor / camera 106 can be placed above the photography platform 103 and can look straight down onto the photography platform 103. In some embodiments, the field of view of the depth sensor / camera 106 can include at least the entire photography platform 103. Plants, plant identification cards 109, pucks, and a color calibration card 110 can be placed on the photography platform 103 so that their contents are captured each time a photo is taken by the depth sensor / camera 106.

[0026] The calibrated light source 102 can be mounted above the photography platform 103 and can illuminate straight down onto the photography platform 103. The calibrated light source 102 may be commercial, off-the-shelf LED panel lights or any other light source which produces consistent color temperatures. The calibrated light source 102 can be bright enough so that it produces a significant portion of the light on the photography platform 103. In some embodiments, environmental lighting can provide the remaining light. The calibrated light source 102 can help remove shadows and provide a more consistent color temperature for accurate color measurements.

[0027] While embodiments of the invention include a depth sensor / camera 106 for capturing images of plants in the visible light spectrum, those skilled in the art will appreciate that various sensors may be used in concert with, or instead of, the depth sensor / camera 106 to capture images of plants in the infrared and / or ultraviolet spectra.

[0028] Embodiments of the invention can use plant identification cards 109 that define the method to be used for identifying the plant being analyzed. This may be through the use of:• A product card that specifically identifies an individual plant;• A trial card that identifies which trial this measurement is associated with; and / or• A lot barcode that identifies a group of plants. Lot barcodes can be used when the plants being studied are small in size and numerous. In such cases, individually identifying each plant would be tedious.

[0029] Embodiments of the invention may include a button panel 107 for user input. The button panel 107 may be a commercial, off-the-shelf product such as the Elgato Stream Deck (https: / / www.elgato.eom / us / en / s / welcome-to-stream-deck) or other similar interface. In some embodiments, a button panel 107 comprising individual liquid crystal displays (LCDs) for each button can be used to allow altering the display of each button. In embodiments, an “Undo” button can appear for a short time after a capture. In some embodiments, after an elapsed time, the Undo button can be removed from the button panel 107 by turning off the LCD display for that button. In some embodiments, other less sophisticated button systems can also be used, for example much of the same functionality can be achieved with simple light-up buttons with physical labels.

[0030] Embodiments of the invention may use a barcode scanner for situations in which QR codes are not available, such as with a product barcode or lot code.

[0031] Embodiments of the invention can use a display 105 that can be any type of computer display. The display 105 can contain status information that, in embodiments, can include about four or more boxes that show that the necessary pucks and barcodes in the image have been recognized. In some embodiments, there may not be a display 105 and the information may be conveyed to the user via a series of labels and lights to denote the status.Calibration

[0032] Some of the measuring devices require calibration to ensure proper results. Generally, calibration should be done at least once a day, prior to making any measurements. Calibration allows the system to ignore such things as stains on the photography platform and provide more accurate results.

[0033] To calibrate, the following steps can be performed in some embodiments:• Removing any objects, configuration pucks, and detritus from the photography platform 103;• Cleaning and wiping down the photography platform 103 with a soft microfiber or similar cloth;• Placing the configuration pucks 108 back on the photography platform 103 and pressing a “Calibrate” button on button panel 107.

[0034] After calibration, the system can be considered ready to begin collecting data. Periodic calibration can be performed during heavy usage following the same steps above.

[0035] During the calibration step, a subtraction mask can be created which is used to subtract the background from future plant images. The scale 101 can be tared when possible.Pucks

[0036] The photography platform 103 itself can be designed to be largely agnostic to what is being measured. Virtually any object can be recorded if it fits on the photography platform 103. However, for that recording to be analyzed appropriately and recorded in the cloud, the collection itself can be sent along with a few pieces of information. In some embodiments, the instructions of how this image should be analyzed can be communicated via configuration pucks 108.

[0037] Configuration pucks 108 can specify how the product / crop should be prepared to be accurately captured and how the number puck should be interpreted. The number puck’s purpose changes based on the configuration puck 108 that is in use, i.e. , placed on the photography platform 103. For some crops such as lettuce, it may be difficult or impossible to fit all of the leaves of the plant on the photographyplatform at the same time. As such, the first set of leaves can be placed on the photography platform 103 with puck number 1 , and then the subsequent leaves can be placed on the photography platform 103 with puck number 2, repeating this process until all leaves are captured. In some embodiments, in each capture, the configuration puck 108 and the number puck can be placed in their respective designated spots on the photography platform 103 for the capture to be successful.Plant Identification Card

[0038] The next piece of required data is the plant identification card 109. One of two methods can be used to include this data. First, an attached 2D barcode scanner that can be used if the plant identification card 109 contains a 2D barcode. When the item has been scanned correctly, the capture can proceed as normal. Second, the plant identification card 109 itself can be physically placed on the designated spot on the photography platform 103, and the code can be scanned by the kiosk when a capture occurs.Color Calibration Card

[0039] Embodiments of the invention also include a color calibration card 110, such as the Datacolor SpyderCHECKR - 24 color card(https: / / www.datacolor.com / spyder / welcome-to-spyder-checkr / ). The color calibration card 110 is a tool that helps ensure accurate color reproduction when taking a photo. The color calibration card 110 contains a number of color swatches of known values. By photographing these known values, a white balance adjustment can be applied later to correct any shifts in the color representations in the photo caused by changes in the lighting conditions. The use of the color calibration Card 110 is important because the platform is mobile and the lighting environment the platform is in may change as the platform is moved around during collection of the data.

[0040] There is a designated spot on the photography platform 103 for color calibration card 110.Capture

[0041] In embodiments of the invention, when the configuration puck 108, number puck, plant identification card 109, and color calibration card 110 are correctly placedor inputted into the computer, all of the boxes in the display are green, and the capture can occur. If any red boxes are still in the display, the required item has not been inputted or is not located in within the designated spot in the frame.

[0042] If all the boxes are green, pressing the capture button on the button panel 107 performs the capture and sends the data to the cloud for processing and aggregation.

[0043] Those skilled in the art will appreciate that the various display elements, e.g., boxes, may be any desired color for any desired function.

[0044] An associated metadata file with the QR codes is generated along with the color and depth images. The QR codes being visible in the image also allows them to be reprocessed later in the event the metadata file is lost.Undo

[0045] For a short period of time after a capture has occurred, an Undo button appears on the button panel 107. When this button is pressed by the user, it deletes the last capture from both the local device as well as in the cloud. This action is useful if it is discovered after the fact that an item is missing from the capture or otherwise obscured.Configuration Puck Examples

[0046] Examples of the configuration pucks 108 are shown in Figures 2-24, in which Figures 2-8 provide information for leafy produce, such as whole head and leafy lettuce, Figures 9-11 provide information for stemmy, leafy produce such as kale, Figures 12 and 13 provide information for melons, Figures 14 and 15 provide information for strawberries, Figures 16 and 17 provide information for pea shoots, Figure 18 provides information for peppers, Figure 19 provides information for tomatoes, Figure 20 provides information for green onions, and Figures 21-24 provide information for Chinese celery.

[0047] For each class of produce there may be more than one configuration puck 108. For example, while Figures 2-8 provide information for leafy produce, such as whole head and leafy lettuce, Figure 2 concerns whole head lettuce, Figure 3 a top down view of whole head lettuce, Figure 4 concerns individual leaves of a whole headlettuce, Figure 5 concerns a whole head of lettuce that has been bisected, Figure 6 concerns portions of a leafy lettuce that are marketable, Figure 7 concerns offcuts and the core of the leafy lettuce, which are not marketable, and Figure 8 concerns a bisected whole head of a leafy lettuce.

[0048] The format of each configuration puck 108 is preferably consistent. Taking Figure 2 as an example, each configuration puck includes a name 200, an icon 201 , a QR code 202, an expected platform setup 203, number puck usage information 204, a list of measurements gathered 205, and a system code 206. The name 200 and the icon 201 allow an operator of the invention to identify, by name and visually, respectively, the plant being processed. In some embodiments, the name 200 can be a human-friendly or conventional name of the plant. In some embodiments, the system code 206 can be an identifier of the plant such as, for example, an alphanumeric code, or any kind of unique identifier for the plant. In some embodiments, the QR code 202 can represent, as a two-dimensional bar code, at least one element of information provided on the configuration puck 108. In some embodiments, the system code 206 can be a scientific name of the plant. In some embodiments, the expected platform setup 203 can provide instructions for preparing the plant for capture. For example, when the plant is a leafy vegetable, the expected platform setup 203 can include an instruction to remove a plug attached to the leafy section of the plant and lay the plant on a side of the plant. In some embodiments, when the plant is, for example, a strawberry, the expected platform setup 203 can include an instruction to place the strawberry in a strawberry holder (not shown) with a cut side of the strawberry facing up and perpendicular to a view of the camera. In some embodiments, the number puck usage information 204 can include an instruction of when to use the same number puck and when to use a different number puck for a part of the plant. For example, when the plant is lettuce, the number puck usage information 204 can include an instruction to use the same number puck for multiple captures of the same head of lettuce but a different number puck for each unique head of the lettuce. In some embodiments, the list of measurements gathered 205 can include at least one measurement to identify the phenotype of the plant. For example, when the plant is a leafy vegetable, the list of measurements gathered 205 can include a head height, a head width, a head weight, and a head color of the plant.

[0049] Those skilled in the art will appreciate that more or fewer fields and / or different fields can be placed on the configuration pucks 108.

[0050] In embodiments of the invention, all of the pucks are a two-layer plastic, white over black, that have been laser etched. They are stored in a drawer under the photography platform 103. The plant identification card 109 discussed above is stored with the plant. The configuration pucks 108 all remain on the cart.Cloud Processing

[0051] The raw data collected at the point of capture is uploaded to the cloud, which can be done over FTP, HTTP, or other common media transfer protocols. Information captured by the configuration pucks 108 provide the cloud system with enough information to determine the computer vision processing necessary for that particular configuration. This system determines from the configuration puck 108 what metadata needs to be stored with the image and what image processing algorithm is required to produce the relevant metrics. The system then uses key identifiers embedded in the plant identification puck, also referred to herein as plant identification card 109, to join to relevant cloud-based databases to obtain required metadata, such as variety, or production lot, etc. The system then uses information embedded in the configuration pucks 108 to determine the appropriate algorithm or set of algorithms to invoke to produce the required measurements. These measurements and their related metadata are then saved to files in the cloud, i.e. , in bucket storage, or can be entered into an existing database.Computer Vision Algorithms

[0052] Each measurement to be acquired from images requires a bespoke algorithm which translates the image data into a specified measurement. This suite of algorithms is part of the cloud processing system. Algorithms contain some combination of basic data transformations, such as color-correction and machine learning models (CNNs, etc.) that take images as input and produce a desired metric. These algorithms are developed in advance using training datasets and may or may not use existing libraries.

[0053] In some embodiments, the system can determine, based on instructions provided on the configuration pucks 108, which computer vision algorithm is best suitedfor processing the image of a particular plant and invoke that computer vision algorithm to process the plant image. For example, a computer vision algorithm may be optimized to process images of leafy vegetables by distinguishing between various shades of green and assigning a numerical value for each shade. Other computer vision algorithms can be optimized to process images of plants of other colors. Another computer vision algorithm can be optimized to analyze a plant image to determine a size, shape, or surface area of a plant. Another computer vision algorithm may be optimized to process plant images with well-defined shapes comprising sharp edges. Yet another computer vision algorithm can be optimized to process images of plants with thin and fuzzy leaves, such as dill. Yet another computer vision algorithm can be optimized to process images of spherical fruits to determine their circumference or volume. Another computer vision algorithm can be optimized to estimate a length, width, or depth of a plant from the image. In some embodiments, the system can choose multiple computer vision algorithms to process a plant image based on the type of plant and / or instructions provided on the configuration pucks 108. A person having ordinary skill in the art will recognize that various other computer vision algorithms not listed here are possible.

[0054] In some embodiments, the invoked computer vision algorithm can process the plant image and produce at least one measurement of the plant. In some embodiments, the at least one measurement can be a weight, size, leaf, or fruit color, a growth level, a morphology, a structure, or a composition of the plant. In some embodiments, the invoked computer vision algorithm can store the at least one measurement of the plant image, along with metadata associated with the plant image, in a database in the cloud.Computer Implementation

[0055] Figure 25 is a block diagram of a computer system 2500 as may be used to implement certain features of some of the embodiments. The computer system 2500 may be a server computer, a client computer, a personal computer (PC), a user device, a tablet PC, a laptop computer, a personal digital assistant (PDA), a cellular telephone, an iPhone, an iPad, a Blackberry, a processor, a telephone, a web appliance, a network router, switch, or bridge, a console, a hand-held console, a (hand-held) gaming device, a music player, any portable, mobile, hand-held device, wearable device, or anymachine capable of executing a set of instructions (sequential or otherwise) that specify actions to be taken by that machine.

[0056] The computing system 2500 may include one or more central processing units (“processors”) 2505, memory 2510, input / output devices 2525, e.g., keyboard and pointing devices, touch devices, display devices, storage devices 2520, e.g, disk drives, and network adapters 2530, e.g., network interfaces, that are connected to an interconnect 2515. The interconnect 2515 is illustrated as an abstraction that represents any one or more separate physical buses, point to point connections, or both connected by appropriate bridges, adapters, or controllers. The interconnect 2515, therefore, may include, for example, a system bus, a Peripheral Component Interconnect (PCI) bus or PCI-Express bus, a HyperTransport or industry standard architecture (ISA) bus, a small computer system interface (SCSI) bus, a universal serial bus (USB), IIC (12C) bus, or an Institute of Electrical and Electronics Engineers (IEEE) standard 1394 bus, also called Firewire.

[0057] The memory 2510 and storage devices 2520 arc computer-readable storage media that may store instructions that implement at least portions of the various embodiments. In addition, the data structures and message structures may be stored or transmitted via a data transmission medium, e.g., a signal on a communications link. Various communications links may be used, e.g., the Internet, a local area network, a wide area network, or a point-to-point dial-up connection. Thus, computer-readable media can include computer-readable storage media, e.g., non-transitory media, and computer-readable transmission media.

[0058] The instructions stored in memory 2510 can be implemented as software and / or firmware to program the processor 2505 to carry out actions described above. In some embodiments, such software or firmware may be initially provided to the processing system 2500 by downloading it from a remote system through the computing system 2500, e.g., via network adapter 2530.

[0059] The various embodiments introduced herein can be implemented by, for example, programmable circuitry, e.g., one or more microprocessors, programmed with software and / or firmware, or entirely in special-purpose hardwired (non-programmable) circuitry, or in a combination of such forms. Special-purpose hardwired circuitry may be in the form of, for example, one or more ASICs, PLDs, FPGAs, etc.

[0060] Figure 26 is a block diagram that illustrates an example artificial intelligence (Al) system 2600 that can implement aspects of the present technology. In some embodiments, the accuracy of one or more computer vision algorithms for performing measurements of the various physical characteristics of a plant can be improved by training the one or more computer vision algorithms based on Al system 2600. In some embodiments, the one or more computer vision algorithms can be trained to perform measurements by feeding a training dataset to the Al system 2600. In some embodiments, the Al system 2600 can be trained to acquire, process, analyze, and understand an image of the plant captured by the system. The Al system 2600 can be trained prior to using the disclosed technology to evaluate plant phenotype in a realtime or near real-time operating environment. In some embodiments, training the Al system 2600 can comprise feeding images of plants having various physical characteristics to the Al system 2600 and running the Al model 2630 on the images. In some embodiments, the Al system 2600 can be implemented in a cloud computing environment.

[0061] The Al system 2600 is implemented using components of the example computer system 2700 illustrated and described in more detail with reference to Figure 25. For example, the Al system 2600 can be implemented on the processor 2702 using instructions programmed in the non-volatile memory 2710 illustrated and described in more detail with reference to Figure 27. Likewise, implementations of the Al system 2600 can include different and / or additional components or be connected in different ways. Figure 26 illustrates a layered architecture of Al system 2600 that can implement the Al models or computer vision algorithms of the system 2800 of Figure 28, in accordance with some implementations of the present technology. Accordingly, the system 2800 can include one or more components of the Al system 2600.

[0062] As shown, the Al system 2600 can include a set of layers, which conceptually organize elements within an example network topology for the Al system’s architecture to implement a particular Al model 2630. Generally, an Al model 2630 is a computer-executable program implemented by the Al system 2600 that analyzes data to make predictions. Information can pass through each layer of the Al system 2600 to generate outputs for the Al model 2630. The layers can include a data layer 2602, a structure layer 2604, a model layer 2606, and an application layer 2608. The algorithm 2616 of the structure layer 2604 and the model structure 2620 and model parameters2622 of the model layer 2606 together form an example Al model 2630. The optimizer 2626, loss function engine 2624, and regularization engine 2628 work to refine and optimize the Al model 2630, and the data layer 2602 provides resources and support for application of the Al model 2630 by the application layer 2608.

[0063] The data layer 2602 acts as the foundation of the Al system 2600 by preparing data for the Al model 2630. As shown, the data layer 2602 can include two sub-layers: a hardware platform 2610 (e.g., the computer devices 304a, 304b, server 340, and network 336 described in more detail with reference to Figure 3) and one or more software libraries 2612. The hardware platform 2610 can be designed to perform operations for the Al model 2630 and include computing resources for storage, memory, logic, and networking, such as the resources described in relation to Figure 6. The hardware platform 2610 can process amounts of data using one or more servers. The servers can perform backend operations such as matrix calculations, parallel calculations, machine learning (ML) training, and the like. Examples of servers used by the hardware platform 2610 include central processing units (CPUs) and graphics processing units (GPUs). CPUs are electronic circuitry designed to execute instructions for computer programs, such as arithmetic, logic, controlling, and input / output (I / O) operations, and can be implemented on integrated circuit (IC) microprocessors. GPUs are electric circuits that were originally designed for graphics manipulation and output but may be used for Al applications due to their vast computing and memory resources. GPUs use a parallel structure that generally makes their processing more efficient than that of CPUs. In some instances, the hardware platform 2610 can include computing resources, (e.g., servers, memory, etc.) offered by a cloud services provider. The hardware platform 2610 can also include computer memory for storing data about the Al model 2630, application of the Al model 2630, and training data for the Al model 2630. The computer memory can be a form of random-access memory (RAM), such as dynamic RAM, static RAM, and non-volatile RAM.

[0064] The software libraries 2612 can be thought of as suites of data and programming code, including executables, used to control the computing resources of the hardware platform 2610. The programming code can include low-level primitives (e.g., fundamental language elements) that form the foundation of one or more low- level programming languages, such that servers of the hardware platform 2610 can use the low-level primitives to carry out specific operations. The low-level programminglanguages do not require much, if any, abstraction from a computing resource’s instruction set architecture, allowing them to run quickly with a small memory footprint. Examples of software libraries 2612 that can be included in the Al system 2600 include INTEL Math Kernel Library, NVIDIA cuDNN, EIGEN, and OpenBLAS.

[0065] The structure layer 2604 can include an ML framework 2614 and an algorithm 2616. The ML framework 2614 can be thought of as an interface, library, or tool that allows users to build and deploy the Al model 2630. The ML framework 2614 can include an open-source library, an application programming interface (API), a gradient-boosting library, an ensemble method, and / or a deep learning toolkit that work with the layers of the Al system to facilitate development of the Al model 2630. For example, the ML framework 2614 can distribute processes for application or training of the Al model 2630 across multiple resources in the hardware platform 2610. The ML framework 2614 can also include a set of pre-built components that have the functionality to implement and train the Al model 2630 and allow users to use pre-built functions and classes to construct and train the Al model 2630. Thus, the ML framework 2614 can be used to facilitate data engineering, development, hyperparameter tuning, testing, and training for the Al model 2630. Examples of ML frameworks 2614 that can be used in the Al system 2600 include TENSORFLOW, PYTORCH, SCIKIT-LEARN, KERAS, LightGBM, RANDOM FOREST, and AMAZON WEB SERVICES.

[0066] The algorithm 2616 can be an organized set of computer-executable operations used to generate output data from a set of input data and can be described using pseudocode. In some implementations, the algorithm 2616 can be a computer vision algorithm configured to perform phenotype measurements on a plant image. The algorithm 2616 can include complex code that allows the computing resources to learn from new input data (e.g. , plant images fed to algorithm 2616 for training purposes) and create new / modified outputs based on what was learned. In some implementations, the new / modified outputs can be a set of parameters identifying each distinct plant phenotype in the training data based on various physical characteristics of the plant. In some implementations, the algorithm 2616 can build the Al model 2630 through being trained while running computing resources of the hardware platform 2610. This training allows the algorithm 2616 to make predictions or decisions without being explicitly programmed to do so. Once trained, the algorithm 2616 can run at the computing resources as part of the Al model 2630 to make predictions or decisions, improvecomputing resource performance, or perform tasks. The algorithm 2616 can be trained using supervised learning, unsupervised learning, semi-supervised learning, and / or reinforcement learning.

[0067] Using supervised learning, the algorithm 2616 can be trained to learn patterns (e.g., map input data to output data) based on labeled training data, such as plant images. The training data may be labeled by an external user or operator. For instance, a user may collect a set of training data, such as by capturing data from sensors, images from a camera, outputs from a model, and the like. In an example implementation, training data can include native-format data collected (e.g., in the form of content from a user of computer device 304a) from various source computing systems described in relation to Figure 1. For example, instead of using conventional digital images, the data can include raw analog signal values of captured images and their parameters. The user may label the training data based on one or more classes and train the Al model 2630 by inputting the training data to the algorithm 2616. The algorithm determines how to label the new data based on the labeled training data. The user can facilitate collection, labeling, and / or input via the ML framework 2614. In some instances, the user may convert the training data to a set of feature vectors for input to the algorithm 2616. Once trained, the user can test the algorithm 2616 on new data to determine if the algorithm 2616 is predicting accurate labels for the new data. For example, the user can use cross-validation methods to test the accuracy of the algorithm 2616 and retrain the algorithm 2616 on new training data if the results of the cross-validation are below an accuracy threshold.

[0068] Supervised learning can involve classification and / or regression. Classification techniques involve teaching the algorithm 2616 to identify a category of new observations based on training data and are used when input data for the algorithm 2616 is discrete. Said differently, when learning through classification techniques, the algorithm 2616 receives training data labeled with categories (e.g., classes) such as plant phenotypes and determines how features observed in the training data (e.g., various physical characteristics such as weight, size, leaf or fruit color, growth, morphology, structure, and composition of the plant) relate to the categories (e.g., a lettuce, a strawberry, a leafy vegetable, a fruit, etc.). Once trained, the algorithm 2616 can categorize new data by analyzing the new data for features that map to the categories. Examples of classification techniques include boosting, decision treelearning, genetic programming, learning vector quantization, k-nearest neighbor (k-NN) algorithm, and statistical classification.

[0069] Regression techniques involve estimating relationships between independent and dependent variables and are used when input data to the algorithm 2616 is continuous. Regression techniques can be used to train the algorithm 2616 to predict or forecast relationships between variables. To train the algorithm 2616 using regression techniques, a user can select a regression method for estimating the parameters of the model. The user collects and labels training data that is input to the algorithm 2616 such that the algorithm 2616 is trained to understand the relationship between data features and the dependent variable(s). Once trained, the algorithm 2616 can predict missing historic data or future outcomes based on input data. Examples of regression methods include linear regression, multiple linear regression, logistic regression, regression tree analysis, least squares method, and gradient descent. In an example implementation, regression techniques can be used, for example, to estimate and fill in missing data for machine-learning based pre-processing operations.

[0070] Under unsupervised learning, the algorithm 2616 learns patterns from unlabeled training data. In particular, the algorithm 2616 is trained to learn hidden patterns and insights of input data, which can be used for data exploration or for generating new data. Here, the algorithm 2616 does not have a predefined output, unlike the labels output when the algorithm 2616 is trained using supervised learning. Said another way, unsupervised learning is used to train the algorithm 2616 to find an underlying structure of a set of data, group the data according to similarities, and represent that set of data in a compressed format. In some embodiments, the algorithm 2616 can be trained to determine parameters associated with various physical characteristics of plants from unlabeled training images of plants.

[0071] A few techniques can be used in supervised learning: clustering, anomaly detection, and techniques for learning latent variable models. Clustering techniques involve grouping data into different clusters that include similar data, such that other clusters contain dissimilar data. For example, during clustering, data with possible similarities remain in a group that has less or no similarities to another group. Examples of clustering techniques include density-based methods, hierarchical based methods, partitioning methods, and grid-based methods. In one example, the algorithm 2616 maybe trained to be a k-means clustering algorithm, which partitions n observations in k clusters such that each observation belongs to the cluster with the nearest mean serving as a prototype of the cluster. Anomaly detection techniques are used to detect previously unseen rare objects or events represented in data without prior knowledge of these objects or events. Anomalies can include data that occur rarely in a set, a deviation from other observations, outliers that are inconsistent with the rest of the data, patterns that do not conform to well-defined normal behavior, and the like. When using anomaly detection techniques, the algorithm 2616 may be trained to be an Isolation Forest, local outlier factor (LOF) algorithm, or K-nearest neighbor (k-NN) algorithm. Latent variable techniques involve relating observable variables to a set of latent variables. These techniques assume that the observable variables are the result of training on the latent variables and that the observable variables have nothing in common after controlling for the latent variables. Examples of latent variable techniques that may be used by the algorithm 2616 include factor analysis, item response theory, latent profile analysis, and latent class analysis.

[0072] The model layer 2606 implements the Al model 2630 using data from the data layer and the algorithm 2616 and ML framework 2614 from the structure layer 2604, thus enabling decision-making capabilities of the Al system 2600. The model layer 2606 includes a model structure 2620, model parameters 2622, a loss function engine 2624, an optimizer 2626, and a regularization engine 2628.

[0073] The model structure 2620 describes the architecture of the Al model 2630 of the Al system 2600. The model structure 2620 defines the complexity of the pattern / relationship that the Al model 2630 expresses. Examples of structures that can be used as the model structure 2620 include decision trees, support vector machines, regression analyses, Bayesian networks, Gaussian processes, genetic algorithms, and artificial neural networks (or, simply, neural networks). The model structure 2620 can include a number of structure layers, a number of nodes (or neurons) at each structure layer, and activation functions of each node. Each node’s activation function defines how a node converts data received to data output. The structure layers may include an input layer of nodes that receive input data and an output layer of nodes that produce output data. The model structure 2620 may include one or more hidden layers of nodes between the input and output layers. The model structure 2620 can be an Artificial Neural Network (or, simply, neural network) that connects the nodes in the structuredlayers such that the nodes are interconnected. Examples of neural networks include Feedforward Neural Networks, convolutional neural networks (CNNs), Recurrent Neural Networks (RNNs), Autoencoder, and Generative Adversarial Networks (GANs).

[0074] The model parameters 2622 represent the relationships learned during training and can be used to make predictions and decisions based on input data. For example, the Al model 2630 can learn based on training images of plants fed to the Al system 2600 that a particular plant has a particular range of colors, shapes, or weights, and can hence determine model parameters 2622 to include color, shape, and weight with that plant. The model parameters 2622 can weight and bias the nodes and connections of the model structure 2620. For instance, when the model structure 2620 is a neural network, the model parameters 2622 can weight and bias the nodes in each layer of the neural networks, such that the weights determine the strength of the nodes and the biases determine the thresholds for the activation functions of each node. The model parameters 2622, in conjunction with the activation functions of the nodes, determine how input data is transformed into desired outputs. The model parameters 2622 can be determined and / or altered during training of the algorithm 2616.

[0075] The loss function engine 2624 can determine a loss function, which is a metric used to evaluate the Al model’s performance during training. For instance, the loss function engine 2624 can measure the difference between a predicted output of the Al model 2630 and the actual output of the Al model 2630 and is used to guide optimization of the Al model 2630 during training to minimize the loss function. The loss function may be presented via the ML framework 2614, such that a user can determine whether to retrain or otherwise alter the algorithm 2616 if the loss function is over a threshold. In some instances, the algorithm 2616 can be retrained automatically if the loss function is greater than the threshold. Examples of loss functions include a binarycross entropy function, hinge loss function, regression loss function (e.g., mean square error, or quadratic loss), mean absolute error function, smooth mean absolute error function, log-cosh loss function, and quantile loss function.

[0076] The optimizer 2626 adjusts the model parameters 2622 to minimize the loss function during training of the algorithm 2616. In other words, the optimizer 2626 uses the loss function generated by the loss function engine 2624 as a guide to determine what model parameters lead to the most accurate Al model. Examples ofoptimizers include Gradient Descent (GD), Adaptive Gradient Algorithm (AdaGrad), Adaptive Moment Estimation (Adam), Root Mean Square Propagation (RMSprop), Radial Base Function (RBF) and Limited-memory BFGS (L-BFGS). The type of optimizer 2626 used may be determined based on the type of model structure 2620 and the size of data and the computing resources available in the data layer 2602.

[0077] The regularization engine 2628 executes regularization operations. Regularization is a technique that prevents over- and under-fitting of the Al model 2630. Overfitting occurs when the algorithm 2616 is overly complex and too adapted to the training data, which can result in poor performance of the Al model 2630. Underfitting occurs when the algorithm 2616 is unable to recognize even basic patterns from the training data such that it cannot perform well on training data or on validation data. The optimizer 2626 can apply one or more regularization techniques to fit the algorithm 2616 to the training data properly, which helps constrain the resulting Al model 2630 and improves its ability for generalized application. Examples of regularization techniques include lasso (L1 ) regularization, ridge (L2) regularization, and elastic (L1 and L2 regularization).

[0078] The application layer 2608 describes how the Al system 2600 is used to solve problem or perform tasks. In some embodiments, the application layer 2608 can be implemented in the cloud system. In embodiments, the application layer 2608 can receive a plant image captured by the system, identifying information of the plant, information contained in the configuration puck, and an instruction to evaluate the phenotype of the plant. In response, the application layer 2608 can invoke an appropriate computer vision algorithm suitable for processing that particular type of plant. The invoked computer vision algorithm can extract various characteristics of the plant from the plant image, compare them with model parameters 2622 of the Al model 2630, and determine the plant phenotype based on the comparison. Once the plant phenotype is identified, the application layer 2608 can enter the plant information and the determined phenotype in a database of the cloud system.Computer System

[0079] Figure 27 is a block diagram that illustrates an example of a computer system 2700 in which at least some operations described herein can be implemented. As shown, the computer system 2700 can include: one or more processors 2702, mainmemory 2706, non-volatile memory 2710, a network interface device 2712, video display device 2718, an input / output device 2720, a control device 2722 (e.g., keyboard and pointing device), a drive unit 2724 that includes a storage medium 2726, and a signal generation device 2730 that are communicatively connected to a bus 2716. The bus 2716 represents one or more physical buses and / or point-to-point connections that are connected by appropriate bridges, adapters, or controllers. Various common components (e.g., cache memory) are omitted from Figure 27 for brevity. Instead, the computer system 2700 is intended to illustrate a hardware device on which components illustrated or described relative to the examples of the figures and any other components described in this specification can be implemented.

[0080] The computer system 2700 can take any suitable physical form. For example, the computer system 2700 can share a similar architecture as that of a server computer, personal computer (PC), tablet computer, mobile telephone, game console, music player, wearable electronic device, network-connected (“smart”) device (e.g., a television or home assistant device), AR / VR systems (e.g., head-mounted display), or any electronic device capable of executing a set of instructions that specify action(s) to be taken by the computer system 2700. In some implementations, the computer system 2700 can be an embedded computer system, a system-on-chip (SOC), a single-board computer system (SBC) or a distributed system such as a mesh of computer systems or include one or more cloud components in one or more networks. Where appropriate, one or more computer systems 2700 can perform operations in real time, near real time, or in batch mode.

[0081] The network interface device 2712 enables the computer system 2700 to mediate data in a network 2714 with an entity that is external to the computer system 2700 through any communication protocol supported by the computer system 2700 and the external entity. Examples of the network interface device 2712 include a network adaptor card, a wireless network interface card, a router, an access point, a wireless router, a switch, a multilayer switch, a protocol converter, a gateway, a bridge, bridge router, a hub, a digital media receiver, and / or a repeater, as well as all wireless elements noted herein.

[0082] The memory (e.g., main memory 2706, non-volatile memory 2710, machine-readable medium 2726) can be local, remote, or distributed. Although shownas a single medium, the machine-readable medium 2726 can include multiple media (e.g., a centralized / distributed database and / or associated caches and servers) that store one or more sets of instructions 2728. The machine-readable (storage) medium 2726 can include any medium that is capable of storing, encoding, or carrying a set of instructions for execution by the computer system 2700. The machine-readable medium 2726 can be non-transitory or comprise a non-transitory device. In this context, a non-transitory storage medium can include a device that is tangible, meaning that the device has a concrete physical form, although the device can change its physical state. Thus, for example, non-transitory refers to a device remaining tangible despite this change in state.

[0083] Although implementations have been described in the context of fully functioning computing devices, the various examples are capable of being distributed as a program product in a variety of forms. Examples of machine-readable storage media, machine-readable media, or computer-readable media include recordable-type media such as volatile and non-volatile memory 2710, removable flash memory, hard disk drives, optical disks, and transmission-type media such as digital and analog communication links.

[0084] In general, the routines executed to implement examples herein can be implemented as part of an operating system or a specific application, component, program, object, module, or sequence of instructions (collectively referred to as “computer programs”). The computer programs typically comprise one or more instructions (e.g. , instructions 2704, 2708, 2728) set at various times in various memory and storage devices in computing device(s). When read and executed by the processor 2702, the instruction(s) cause the computer system 2700 to perform operations to execute elements involving the various aspects of the disclosure.Example System

[0085] Figure 28 is a block diagram that illustrates an example of a system 2800 in which at least some aspects of the present technology are implemented. In some implementations, the system 2800 can be configured to streamline the process of gathering and analyzing plant data to enable accurate and objective plant phenotyping measurements while eliminating human subjectivity in the data collection process. Some embodiments can comprise a mobile cart system 2802, including a photographyplatform 2804 with a uniform color surface, an overhead-mounted camera and depth sensor system 2806, a scale 2808 for weight measurements, and a calibrated light source 2810 configured to provide consistent color temperature for accurate color measurements of products. In some embodiments, configuration pucks 2812 can include instructions for preparing the product for accurate data capture. In some embodiments, plant identification cards 2814 can include the method to be used for identifying the plant being analyzed. Some embodiments can include color calibration cards 2816 to ensure accurate color reproduction. The scale 2808 can be coupled with the computer system to prevent manual data entry errors and to enable automatic weight recording of the product. The configuration pucks 2812 can specify how products should be prepared for accurate capture and how number pucks should be interpreted. In some embodiments, plant identification cards 2814 can contain a barcode or a QR code. The plant identification card can be physically placed on the photography platform 2804 to ensure proper tracking and documentation of each product. The color calibration card 2816 can contain known color values that enable white balance adjustment to correct for any shifts in color representation caused by varying lighting conditions.

[0086] Figure 29 is a block diagram of a process 2900 in which at least some aspects of the disclosed technology are implemented. The process 2900 can be implemented in a system. At 2902, an image of the product can be captured. At 2904, the captured image of the product can be processed by a local computer of the system. In some embodiments, during local computer processing, the system can interface with the various equipment components, collecting data and generating metadata files that include QR codes and other relevant information. At 2906, the system can generate a metadata file of the captured image. In some embodiments, at 2908, the captured images or the metadata files can be uploaded to a cloud-based system. The cloudbased system can provide scalable computing resources and specialized services for handling complex image processing tasks. At 2910, the system can process the uploaded data using computer vision algorithms. In some embodiments, the computer vision algorithm processing can include performing computer vision tasks that include methods for acquiring, processing, analyzing, and understanding digital images, and extraction of high-dimensional data from the product image to produce numerical or symbolic information. At 2912, the computer vision algorithms can generatemeasurement data associated with the captured images or metadata files. In some embodiments, the measurement data can include various physical characteristics such as weight, size, leaf or fruit color, growth, morphology, structure, and composition of the product. At 2914, the generated measurement data can be stored in a cloud storage or database.

[0087] In some embodiments, the process 2900 can convert the captured optical image of the plant into a digital color code by extracting a color representative of the plant from the optical image. In some embodiments, the color representative of the plant can be a dominant color of the plant. In some embodiments, the digital color code can be represented by a hexadecimal number. In some embodiments, the digital color code can include at least one digital color subcode associated with a level of red, green, blue, cyan, magenta, yellow, black, or white color in the optical image. In some embodiments, the process 2900 can measure a depth or a height of the plant from the captured depth image. In some embodiments, the process 2900 can calculate a surface area or an area of a projection of the plant on the photography platform from the optical image. Thus, the process 2900 converts physical attributes of the plant that are otherwise susceptible to subjective and varying interpretation when made using a manual process into accurate, objective digital and numerical representations made with sophisticated sensors and computing equipment. When the disclosed technology is implemented, the tedious, slow, and error-prone phenotyping of plants by human minds and sensory organs is avoided.

[0088] Figure 30 shows a flowchart of a process 3000 in which aspects of the disclosed technology are implemented. At 3002, a phenotype evaluation platform configured to evaluate a phenotype of a plant is provided. In some embodiments, the phenotype evaluation platform can include a scale disposed in the evaluation platform, a photography platform positioned above the scale and configured to receive the plant, an optical sensor and a depth sensor each positioned above the photography platform, and a calibrated light source positioned above the photography platform. In some embodiments, the optical sensor can be a camera configured to operate in a visible spectrum of light. In some embodiments, the optical sensor can be a camera configured to operate in an infrared or an ultraviolet spectrum of light. At 3004, a configuration puck, a plant identification card, and a color calibration card can be positioned on the photography platform within a field of vision of the optical sensor. In some embodiments,the plant identification card can include an identifier of the plant. In some embodiments, the color calibration card can include at least one color swatch of a known value. In some embodiments, the configuration puck can include an identifier of the plant type, a QR code, and an instruction to position the plant on the photography platform for capturing an image of the plant. At 3006, the plant can be positioned on the photography platform for phenotype evaluation according to the at least one instruction included in the configuration puck. At 3008, the plant can be illuminated with light from the calibrated light source. At 3010, an optical image of the plant can be captured by the optical sensor. At 3012, a depth image of the plant can be captured by the depth sensor. At 3014, a metadata file can be generated based on the captured optical and depth images. In some embodiments, the metadata file can include the QR code and the identifier of the plant. At 3016, a digital color code representing a color of the plant can be extracted. At 3018, a depth of the plant can be calculated from the depth image. At 3020, a phenotype of the plant can be determined based at least on the digital color code and the depth. In some embodiments, determining the phenotype can further be based on a size, a morphology, an architecture, or a composition of the plant. In some embodiments, determining the phenotype can further be based on a weight of the plant as measured by the scale. At 3022, at least one of the optical image, the depth image, the digital color code representing a color of the plant, the depth of the plant, the phenotype of the plant, the QR code, and the identifier of the plant can be displayed on a computer display coupled to the evaluation platform.Remarks

[0089] The above description and drawings are illustrative and are not to be construed as limiting. Numerous specific details are described to provide a thorough understanding of the disclosure. However, in certain instances, well-known details are not described to avoid obscuring the description. Further, various modifications may be made without deviating from the scope of the embodiments.

[0090] Reference in this specification to “one embodiment’’ or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the disclosure. The appearances of the phrase “in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternativeembodiments mutually exclusive of other embodiments. Moreover, various features are described which may be exhibited by some embodiments and not by others. Similarly, various requirements are described which may be requirements for some embodiments but not for other embodiments.

[0091] The terms used in this specification generally have their ordinary meanings in the art, within the context of the disclosure, and in the specific context where each term is used. Certain terms that are used to describe the disclosure are discussed above, or elsewhere in the specification, to provide additional guidance to the practitioner regarding the description of the disclosure. For convenience, certain terms may be highlighted, for example using italics and / or quotation marks. The use of highlighting has no influence on the scope and meaning of a term; the scope and meaning of a term is the same, in the same context, whether or not it is highlighted. It will be appreciated that the same thing can be said in more than one way. One will recognize that “memory” is one form of a “storage” and that the terms may on occasion be used interchangeably.

[0092] Consequently, alternative language and synonyms may be used for any one or more of the terms discussed herein, nor is any special significance to be placed upon whether or not a term is elaborated or discussed herein. Synonyms for certain terms are provided. A recital of one or more synonyms does not exclude the use of other synonyms. The use of examples anywhere in this specification including examples of any term discussed herein is illustrative only and is not intended to further limit the scope and meaning of the disclosure or of any exemplified term. Likewise, the disclosure is not limited to various embodiments given in this specification.

[0093] Without intent to further limit the scope of the disclosure, examples of instruments, apparatus, methods and their related results according to the embodiments of the present disclosure are given above. Note that titles or subtitles may be used in the examples for convenience of a reader, which in no way should limit the scope of the disclosure. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. In the case of conflict, the present document, including definitions will control.

Claims

CLAIMSI / We claim:1 . A method comprising: providing a phenotype evaluation platform configured to evaluate a phenotype of a plant, wherein the phenotype evaluation platform includes a scale, a photography platform positioned above the scale and configured to receive the plant, an optical sensor and a depth sensor each positioned above the photography platform, and a calibrated light source positioned above the photography platform; positioning a configuration puck, a plant identification card, and a color calibration card on the photography platform within a field of vision of the optical sensor, wherein the plant identification card includes an identifier of the plant, wherein the color calibration card includes at least one color swatch of a known value, and wherein the configuration puck includes an identifier of a plant type of the plant, a QR code, and at least one instruction to position the plant on the photography platform for capturing an image of the plant; positioning the plant on the photography platform for phenotype evaluation according to the at least one instruction included in the configuration puck; illuminating the plant with light from the calibrated light source; capturing, by the optical sensor, an optical image of the plant; capturing, by the depth sensor, a depth image of the plant; generating a metadata file based on the captured optical and depth images, wherein the metadata file includes the QR code, the identifier of the plant, and the identifier of the plant type of the plant; extracting, from the optical image, a digital color code representing a color of the plant; calculating, from the depth image, a depth of the plant;determining a phenotype of the plant based at least on the digital color code and the depth; and displaying, on a computer display coupled to the evaluation platform, at least one of the optical image, the depth image, the digital color code representing a color of the plant, the depth of the plant, the phenotype of the plant, the QR code, the identifier of the plant, and the identifier of the plant type.

2. The method of claim 1 further comprising: determining a phenotype of the plant based on a size, a morphology, an architecture, or a composition of the plant.

3. The method of claim 1 , wherein the optical sensor is a camera configured to operate in a visible spectrum of light.

4. The method of claim 1 further comprising: uploading the captured optical and depth images to a cloud computing server; identifying, by the cloud computing server, a computer vision algorithm to be used for processing the captured optical and depth images based on the at least one instruction provided on the configuration puck; and performing, by the computer vision algorithm, a phenotype measurement of the plant.

5. The method of claim 4, wherein the computer vision algorithm is based on an artificial intelligence or machine learning model.

6. The method of claim 1 further comprising: measuring, with the scale, a weight of the plant; and determining the phenotype of the plant based on the weight of the plant.

7. The method of claim 1 , wherein the optical sensor is a camera configured to operate in an infrared or an ultraviolet spectrum of light.

8. A system for evaluating a phenotype of a plant, the system comprising: at least one hardware processor; and at least one non-transitory memory storing instructions, which, when executed by the at least one hardware processor, cause the system to: capture, by an optical sensor positioned above a photography platform coupled to a phenotype evaluation platform, an image of the plant, wherein the photography platform is positioned above a scale and is configured to receive the plant, wherein the phenotype evaluation platform further includes a calibrated light source positioned above the photography platform, wherein the phenotype evaluation platform further includes a configuration puck, a plant identification card, and a color calibration card each positioned on the photography platform within a field of vision of the optical sensor, wherein the configuration puck includes an identifier of a plant type of the plant, and an instruction to position the plant on the photography platform for capturing an image of the plant, and wherein the image is captured by preparing the plant according to at least one instruction provided on the configuration puck; and generate a metadata file based on the captured image, wherein the metadata file includes the identifier of the plant and the identifier of the plant type; and determine a phenotype of the plant based on the captured image.

9. The system of claim 8, wherein the phenotype of the plant is determined based on the captured image by determining a size, a leaf color, a fruit color, a morphology, an architecture, or a composition of the plant.

10. The system of claim 8, wherein the optical sensor is a camera configured to operate in a visible spectrum of light.11 . The system of claim 8 further caused to: upload the captured image to a cloud computing server; identify, at the cloud computing server, a computer vision algorithm to be used for processing the captured image based on the at least one instruction provided on the configuration puck; and perform, using the computer vision algorithm, a phenotype measurement of the plant.

12. The system of claim 11 , wherein the computer vision algorithm is based on an artificial intelligence or machine learning model.

13. The system of claim 8 further caused to: measure, with the scale, a weight of the plant.

14. The system of claim 8, wherein the optical sensor is a camera configured to operate in an infrared or an ultraviolet spectrum of light.

15. The system of claim 8, wherein the optical sensor is a depth camera, and the captured image is a depth image of the plant.

16. A portable apparatus for evaluating phenotype of a plant, the apparatus comprising: a wheeled cart; a scale disposed in the wheeled cart; a photography platform positioned above a scale, wherein the photography platform is configured to receive the plant; an optical sensor positioned above the photography platform; a calibrated light source positioned above the photography platform; a configuration puck positioned within a field of vision of the optical sensor, wherein the configuration puck includes at least one instruction for preparing the plant for phenotype evaluation; a plant identification card positioned within a field of vision of the optical sensor, wherein the plant identification card includes an identifier of the plant; and a color calibration card positioned on the photography platform within a field of vision of the optical sensor, wherein the color calibration card includes at least one color swatch of a known value.

17. The portable apparatus of claim 16 further comprising: a computer display configured to provide a status of the phenotype evaluation, an image of the plant captured by the optical sensor, or a phenotype attribute of the plant measured by the portable apparatus.

18. The portable apparatus of claim 16 further comprising: a button panel configured to control an operation of the optical sensor, the scale, or the calibrated light source.

19. The portable apparatus of claim 16, wherein at least one wheel of the wheeled cart is coupled to an energy source, and wherein the portable apparatus is enabled to be portable under its own power by the at least one wheel coupled to the energy source.

20. The portable apparatus of claim 16, wherein the optical sensor is a camera configured to operate in an infrared or an ultraviolet spectrum of light.

Images