A training machine to generate training data for supervised machine learning. The training machine is a working machine with a camera, a sensor, a body, and a moveable element moveable relative to the body. Training data is generated by capturing images, using the camera of at least a portion of the moveable element; and determining, using the sensor, sensor data which indicates the position of the moveable element in the image. The training data is used to train a machine learning algorithm to determine a position of a moveable element based on an image of at least a portion of the moveable element, without the need for corresponding sensor data.

A training machine to generate training data for supervised machine learning. The training machine is a working machine with a camera, a sensor, a body, and a moveable element moveable relative to the body. Training data is generated by capturing images, using the camera of at least a portion of the moveable element; and determining, using the sensor, sensor data which indicates the position of the moveable element in the image. The training data is used to train a machine learning algorithm to determine a position of a moveable element based on an image of at least a portion of the moveable element, without the need for corresponding sensor data.

A method for automatically controlling a harvesting parameter on a header of a combine harvester. The combine harvester includes the header, a feeder, and a downstream processing device. Crop is cut by the header, transferred to the feeder, and then transported to the processing device. The method includes steps of detecting at least one crop property in the feeder by at least one sensor and outputting at least one corresponding crop property signal; receiving the at least one crop property signal by a control unit; processing the at least one crop property signal in the control unit; transmitting at least one control signal by the control unit to at least one actuator on the header; and executing the at least one control signal in the at least one actuator so as to automatically control the harvesting parameter on the header

A method of harvesting plant product from a plant in a single pass using a combine harvester is disclosed. In the method, the plant has a protein content gradient that varies along a height of the plant. The method includes identifying, along a longitudinally-extending stalk of the plant, an upper protein gradient of the plant including high protein plant product and a lower protein gradient of the plant including lower protein plant product, wherein the high protein plant product from the upper protein gradient of the plant meets a threshold protein content that is higher than that of the lower protein plant product. The method also includes separately and substantially simultaneously harvesting the high protein plant product from the upper protein gradient and the lower protein plant product from the lower protein gradient in the single pass, and isolating the high protein plant product from the lower protein plant product.

A method of harvesting plant product from a plant in a single pass using a combine harvester is disclosed. In the method, the plant has a protein content gradient that varies along a height of the plant. The method includes identifying, along a longitudinally-extending stalk of the plant, an upper protein gradient of the plant including high protein plant product and a lower protein gradient of the plant including lower protein plant product, wherein the high protein plant product from the upper protein gradient of the plant meets a threshold protein content that is higher than that of the lower protein plant product. The method also includes separately and substantially simultaneously harvesting the high protein plant product from the upper protein gradient and the lower protein plant product from the lower protein gradient in the single pass, and isolating the high protein plant product from the lower protein plant product.

A method of controlling a work machine on a worksite includes receiving an indication of a work machine assignment having a worksite location and corresponding assignment criteria associated with completion of the work machine assignment, receiving a set of worksite conditions at the location, and identifying a set of machine capabilities, each machine capability corresponding to operation of a controllable subsystem on the work machine. The method includes generating a likelihood metric indicative of a likelihood that the assignment criteria will be met based on the set of worksite conditions and the set of machine capabilities, comparing the likelihood metric to a threshold, and generating a control signal that controls the work machine based on the comparison.

A method of controlling a work machine on a worksite includes receiving an indication of a work machine assignment having a worksite location and corresponding assignment criteria associated with completion of the work machine assignment, receiving a set of worksite conditions at the location, and identifying a set of machine capabilities, each machine capability corresponding to operation of a controllable subsystem on the work machine. The method includes generating a likelihood metric indicative of a likelihood that the assignment criteria will be met based on the set of worksite conditions and the set of machine capabilities, comparing the likelihood metric to a threshold, and generating a control signal that controls the work machine based on the comparison.

The disclosure relates to a crop merger system for a harvester. The system includes a frame, first and second rollers mounted to the frame, and a belt disposed over the first and second rollers. The system includes a motor operably coupled to and configured to rotate the second roller. The system includes a controller configured to electronically receive as input a desired speed of the belt. The system includes a sensor configured to detect a measured characteristic associated with the crop merger system, and electronically transmit the measured characteristic to the controller. The controller is configured to determine a measured speed of the belt based on the measured characteristic, and if the desired speed and the measured speed are unequal, the controller is configured to regulate the motor to adjust a speed of the belt to be substantially equal to the desired speed.

The disclosure relates to a crop merger system for a harvester. The system includes a frame, first and second rollers mounted to the frame, and a belt disposed over the first and second rollers. The system includes a motor operably coupled to and configured to rotate the second roller. The system includes a controller configured to electronically receive as input a desired speed of the belt. The system includes a sensor configured to detect a measured characteristic associated with the crop merger system, and electronically transmit the measured characteristic to the controller. The controller is configured to determine a measured speed of the belt based on the measured characteristic, and if the desired speed and the measured speed are unequal, the controller is configured to regulate the motor to adjust a speed of the belt to be substantially equal to the desired speed.

Receivers are arranged to detect a corresponding observed phase shift, observed attenuation or other observed signal parameters for its respective microphone. An electronic data processor is adapted to estimate a distribution or quantity of material on the sieve based on the observed phase shift, the observed attenuation or the other observed signal parameters relative to a reference phase shift, a reference attenuation or other reference signal parameter. The operator can be alerted via a user interface if the material on the sieve is unevenly distributed or matches a preestablished distribution profile, or the sieve can be adjusted by an actuator to promote a generally uniform distribution.