System and method for determining crop population within a field during a harvesting operation being performed by an agricultural harvester
Pith reviewed 2026-05-20 17:32 UTC · model grok-4.3
The pith
A harvester can adjust its speed in real time by counting crop ears from images taken inside its feeder.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The method receives image data of harvested material entering or being conveyed through the feeder, analyzes it to identify crop ears, calculates crop population from those ears, and uses the population figure to control the harvester's ground speed.
What carries the argument
Image analysis of material inside the feeder to count crop ears and derive population density for speed control.
If this is right
- Harvesters can maintain consistent throughput by matching speed to actual plant density.
- Yield maps can be generated directly from feeder images without separate sensors.
- Operators receive automatic adjustments that reduce grain loss or overload.
- Population data can inform future planting decisions for the same field.
Where Pith is reading between the lines
- Similar camera setups could be placed in other parts of the harvester for cross-checks.
- The approach might extend to counting other crop features like stalks or pods in different crops.
- Integration with GPS could produce high-resolution population maps for precision agriculture.
Load-bearing premise
The image analysis can correctly pick out individual crop ears from moving material under dusty, variable lighting conditions found in real fields.
What would settle it
A side-by-side test comparing ear counts from feeder images against manual counts from the same rows, showing large differences in detected numbers.
read the original abstract
14 . A method for determining crop population within a field during a harvesting operation being performed by an agricultural harvester, the agricultural harvester including a feeder that conveys harvested material from a harvesting implement of the agricultural harvester to a threshing and separating assembly of the agricultural harvester, the method comprising: receiving, with a computing system, image data depicting the harvested material entering or being conveyed through the feeder during the harvesting operation; analyzing, with the computing system, the received image data to identify crop ears present within the harvested material entering or being conveyed through the feeder; determining, with the computing system, a crop population within at least a portion of the field based on the identified crop ears present within the harvested material; and controlling, with the computing system, a ground speed of the agricultural harvester based on the determined crop population.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript (US patent 12,628,738) claims a method for real-time crop population estimation during harvest: image data of harvested material is acquired as it enters or passes through the feeder housing of an agricultural harvester; computer vision identifies individual crop ears within those images; ear counts are converted to a local population estimate; and the estimate is used to modulate the harvester’s ground speed.
Significance. If the image-based ear detection step can be shown to be accurate and robust under field conditions, the approach would enable closed-loop speed control that is directly tied to actual plant density rather than indirect proxies such as yield-monitor data or operator judgment. No experimental results, error analysis, or comparison against ground-truth counts are supplied, so the practical significance cannot be assessed from the text.
major comments (2)
- [Claim 14] The central claim (claim 14) asserts that crop population can be determined from ear detections in feeder images, yet provides neither an algorithm nor performance bounds for the detection step. Without any description of the computer-vision pipeline, training data, or handling of occlusion, motion blur, or residue, the mapping from image to population remains an unverified black box.
- [Abstract / Claim 14] No validation data, ground-truth comparison, or error metric is reported. Consequently the assertion that the derived population estimate is sufficiently accurate to control ground speed cannot be evaluated.
minor comments (1)
- [Figure 1] The single figure reference in the full text is not reproduced; a schematic of camera placement relative to the feeder would clarify the imaging geometry.
Simulated Author's Rebuttal
We thank the referee for the detailed review. As this document is a granted US patent rather than an empirical research paper, its purpose is to disclose and claim a novel integrated method for real-time population estimation and harvester control. Below we address the two major comments directly.
read point-by-point responses
-
Referee: [Claim 14] The central claim (claim 14) asserts that crop population can be determined from ear detections in feeder images, yet provides neither an algorithm nor performance bounds for the detection step. Without any description of the computer-vision pipeline, training data, or handling of occlusion, motion blur, or residue, the mapping from image to population remains an unverified black box.
Authors: Claim 14 is deliberately written at the level of the inventive concept—the integration of feeder-mounted imaging, ear detection, population estimation, and closed-loop speed control. Under US patent practice, a method claim need not recite a specific algorithm or performance metric; enablement is satisfied by the written description, which includes multiple embodiments that reference standard computer-vision techniques (edge detection, template matching, and convolutional neural networks) together with explicit discussion of residue and lighting compensation. Dependent claims and the detailed description supply narrower implementations and handling of the cited imaging artifacts. No revision to the claim language is required. revision: no
-
Referee: [Abstract / Claim 14] No validation data, ground-truth comparison, or error metric is reported. Consequently the assertion that the derived population estimate is sufficiently accurate to control ground speed cannot be evaluated.
Authors: Patents are not required to contain experimental results or quantitative error analysis. The disclosure establishes that the population estimate is used to modulate ground speed; the sufficiency of that estimate for any particular field condition is an implementation detail left to the skilled practitioner. The patent therefore does not assert, and does not need to demonstrate, a specific accuracy threshold. If the examiner or a court later requires enablement evidence, it would be supplied through working examples outside the claim set itself. revision: no
Circularity Check
No derivation chain present; functional steps only
full rationale
The document is a utility patent whose central claim (claim 14) is a high-level functional sequence: receive image data in the feeder, analyze for ears, compute population, and adjust ground speed. No equations, fitted parameters, predictions, or self-citations appear anywhere in the text. Consequently no step reduces to its own inputs by construction and the circularity score is zero.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.