Abstract: Discussed herein are devices, systems, and methods for decentralized device management. A method can include receiving a first message from a second device, implementing a first machine learning (ML) model that operates on the received first message and an observation to determine a next objective to be completed by the first device, and training a simulator to produce the first message based on the observation.

Abstract: Discussed herein are devices, systems, and methods for beam generating device management. A method can include producing a directed electromagnetic beam, receiving a first message from a second beam generating device, and implementing a first machine learning (ML) model that operates on the received first message to determine a next objective to be completed by the beam generating circuitry.

Abstract: Discussed herein are devices, systems, and methods for improved trajectory planning. A method can include providing two of (i) a first value indicating a change in velocity to alter an orbit of a first object to a transfer orbit; (ii) a second value indicating a range between the first object and a second object; or (iii) a third value indicating an altitude of the first object relative to a celestial body around which the first and second objects are orbiting as input to a machine learning (ML) model, receiving, from the ML model, a holdout value, the holdout value a prediction of the value, of the first value, the second value, and the third value, that was not provided to the ML model, and providing the holdout value to an orbital planner.

Abstract: Disclosed are methods for handwriting recognition. In some aspects, an image representing a page of a sample document is analyzed to identify a region having indications of handwriting. The region is analyzed to determine frequencies of a plurality of geometric features within the region. The frequencies may be compared to profiles or histograms of known language types, to determine if there are similarities between the frequencies in the sample document relative to those of the known language types. In some aspects, machine learning may be used to characterize the document as a particular language type based on the frequencies of the geometric features.

Abstract: Discussed herein are devices, systems, and methods for autonomous, dynamic navigation of scenarios. A method can include implementing a path generation machine learning (ML) technique to determine paths between a device and a goal, determining a node of the paths as an intersection of at least two of the paths, and implementing an executive ML technique to determine which of the at least two paths to take at a node of the graph to reach the goal.

Abstract: A method for determining chlorophyll content of a plant comprises capturing a first image comprising light transmitted through a leaf of a plant; capturing a second image comprising light reflected from the leaf of the plant; estimating, from a plurality of pixels in the first image, a transmissive chlorophyll concentration value of the leaf; estimating a reflectance chlorophyll concentration value for the leaf from a plurality of pixels in the second image using bidirectional reflectance parameters for which a variance of the reflectance chlorophyll concentration value across the plurality of pixels in the second image is reduced; and determining an estimated chlorophyll concentration value for the plant based at least on the transmissive chlorophyll concentration value and the reflectance chlorophyll concentration value.

Abstract: Disclosed are methods for handwriting recognition. In some aspects, an image representing a page of a sample document is analyzed to identify a region having indications of handwriting. The region is analyzed to determine frequencies of a plurality of geometric features within the region. The frequencies may be compared to profiles or histograms of known language types, to determine if there are similarities between the frequencies in the sample document relative to those of the known language types. In some aspects, machine learning may be used to characterize the document as a particular language type based on the frequencies of the geometric features.

Abstract: Methods and apparatus are provided for determining moisture content of corn. In one example, a method includes processing a captured image of an ear of corn using a threshold value to create a segmented binary image comprising a first plurality of blobs, determining at least one characteristic of a plurality of corn kernels represented by the first plurality of blobs, and estimating a moisture value for the ear of corn based on the at least one characteristic of the plurality of corn kernels. In some embodiments, the method includes generating a luminance intensity profile across a region of the captured image containing at least one corn kernel in the plurality of corn kernels, computing a derivative of the luminance intensity profile; an determining, with reference to the derivative of the luminance intensity profile, a location of a boundary of the at least one corn kernel.

Abstract: Methods and apparatus are provided for identifying optimal threshold values to generate optimal binary images from a digital image. In one example, a method includes, for each of a plurality of candidate threshold values, processing an image using the candidate threshold value to create a first segmented image of a first plurality of blobs; for each of the plurality of candidate threshold values, calculating a threshold value evaluation metric as a count of qualifying blobs in the first plurality of blobs in the first segmented image, wherein the qualifying blobs have a fill factor exceeding a predefined fill factor and have an area within a predefined area range; identifying an optimal threshold value from the plurality of candidate threshold values for which the threshold value evaluation metric is maximized; and processing the image using the optimal threshold value to create a second segmented image of a second plurality of blobs.

Abstract: A method for generating a yield estimate for a crop of corn includes capturing a digital image of an ear of corn; processing the digital image of an ear of corn to determine a plurality of dimensions for each of a plurality of kernels on the ear of corn; estimating, from the plurality of dimensions, an average kernel volume for the ear of corn; determining, from the average kernel volume and an estimated kernel density, an average kernel weight for the ear of corn; and estimating, from the average kernel weight, a kernels-per-bushel for the ear of corn.

Abstract: Methods and apparatus for producing multi-view 3D renderings of a scene, and collaboratively sharing information related to the renderings to facilitate producing synthetic viewpoints of the scene. In one example, a method of producing a 3D synthetic viewpoint of the scene includes receiving an image of the scene from a local camera, analyzing the image to determine a context of the scene and to extract at least one surface feature of the scene, developing an initial set of planar world coordinates for the image based on the at least one surface feature to estimate an initial local perspective projection matrix, and mapping image pixels of the image to a world plane to produce the 3D synthetic viewpoint of the scene by using the initial local perspective projection matrix to map image space coordinates to world coordinates.

Abstract: Generally discussed herein are systems, apparatuses, and methods for providing contextually relevant augmented reality (AR) to a wearable device. In one or more embodiments, a method can include extracting features of one or more objects in a location in a field of view of a camera of a wearable device, retrieving a three dimensional model of the location based on the extracted features, assessing a situation of a user associated with the wearable device, modifying the three dimensional model based on the assessed user situation, and presenting, using the wearable device, at least a portion of the modified three dimensional model.

Abstract: A method for generating a yield estimate for a crop of corn includes capturing a digital image of an ear of corn; processing the digital image of an ear of corn to determine a plurality of dimensions for each of a plurality of kernels on the ear of corn; estimating, from the plurality of dimensions, an average kernel volume for the ear of corn; determining, from the average kernel volume and an estimated kernel density, an average kernel weight for the ear of corn; and estimating, from the average kernel weight, a kernels-per-bushel for the ear of corn.

Abstract: A method for determining chlorophyll content of a plant comprises capturing a first image comprising light transmitted through a leaf of a plant; capturing a second image comprising light reflected from the leaf of the plant; estimating, from a plurality of pixels in the first image, a transmissive chlorophyll concentration value of the leaf; estimating a reflectance chlorophyll concentration value for the leaf from a plurality of pixels in the second image using bidirectional reflectance parameters for which a variance of the reflectance chlorophyll concentration value across the plurality of pixels in the second image is reduced; and determining an estimated chlorophyll concentration value for the plant based at least on the transmissive chlorophyll concentration value and the reflectance chlorophyll concentration value.

Abstract: Methods and apparatus are provided for determining moisture content of corn. In one example, a method includes processing a captured image of an ear of corn using a threshold value to create a segmented binary image comprising a first plurality of blobs, determining at least one characteristic of a plurality of corn kernels represented by the first plurality of blobs, and estimating a moisture value for the ear of corn based on the at least one characteristic of the plurality of corn kernels. In some embodiments, the method includes generating a luminance intensity profile across a region of the captured image containing at least one corn kernel in the plurality of corn kernels, computing a derivative of the luminance intensity profile; an determining, with reference to the derivative of the luminance intensity profile, a location of a boundary of the at least one corn kernel.

Abstract: Methods and apparatus are provided for identifying optimal threshold values to generate optimal binary images from a digital image. In one example, a method includes, for each of a plurality of candidate threshold values, processing an image using the candidate threshold value to create a first segmented image of a first plurality of blobs; for each of the plurality of candidate threshold values, calculating a threshold value evaluation metric as a count of qualifying blobs in the first plurality of blobs in the first segmented image, wherein the qualifying blobs have a fill factor exceeding a predefined fill factor and have an area within a predefined area range; identifying an optimal threshold value from the plurality of candidate threshold values for which the threshold value evaluation metric is maximized; and processing the image using the optimal threshold value to create a second segmented image of a second plurality of blobs.

Abstract: Generally discussed herein are systems, apparatuses, and methods for providing contextually relevant augmented reality (AR) to a wearable device. In one or more embodiments, a method can include extracting features of one or more objects in a location in a field of view of a camera of a wearable device, retrieving a three dimensional model of the location based on the extracted features, assessing a situation of a user associated with the wearable device, modifying the three dimensional model based on the assessed user situation, and presenting, using the wearable device, at least a portion of the modified three dimensional model.

Abstract: Communicating between a plurality of nodes includes scanning a set of frequencies to determine spectrum awareness data for one or more frequencies; determining a maximum transmit power for one or more of the frequencies based at least in part on the spectrum awareness data for the frequencies; determining one or more channels, each channel occupying at least a portion of the set of frequencies; and selecting one or more of the one or more channels as a bearer.

Abstract: In one embodiment, a system for identifying a communication device includes a signal generator coupled to a transmit horn and a computing system coupled to a receive horn through a receiver. The signal generator is operable to generate an excitation waveform from the transmit horn such that the communication device passively reflects a response waveform. The computing system is operable to receive the response waveform from the communication device and compare the response waveform to a plurality of reference waveforms to determine the identity of the communication device.

Abstract: Communicating between a plurality of nodes includes: at one or more of the plurality of nodes, sensing spectrum activity on at least one channel; creating and maintaining at least one spectrum awareness table based at least in part on the sensed spectrum activity, determining at least one selected bearer based on the at least one spectrum awareness table; performing an adaptive control channel initialization operation to detect zero or more neighbor nodes.