Detection and Mapping of Kochia Plants and Patches Using High-Resolution Ground Imagery and Satellite Data: Application of Machine Learning
Bibliographic record
Abstract
The presence of Kochia weed can be harmful to crop production. It can grow well in harsh conditions and resist popular herbicides such as glyphosate. It causes chemical stress to crops and spreads quickly, forming large patches. Detecting Kochia early is crucial for effective control. However, it can be hard to identify it in its early stages as it closely resembles early-stage crops. Moreover, Kochia seeds can spread from neighbouring farms through water or wind, making early detection at both field and regional levels necessary. Currently, object-based detection methods are used for Kochia detection at the field level, but there is still a lack of literature on mapping Kochia at a regional level. Our research proposes a methodology for accurately detecting, localizing and quantifying Kochia plants in fields using high-resolution RGB imagery. We also explore the potential of detecting Kochia patches at a regional level using satellite imagery. Our approach uses semantic segmentation techniques to process geotagged RGB images, allowing us to identify and quantify individual Kochia plants in the field. To ensure accurate detection, we have established a minimum Kochia density threshold based on the density of Kochia in RGB images. This threshold enables us to distinguish the spectral signature of satellite imagery pixels with a high density of Kochia. We label the satellite imagery based on the geo-locations where Kochia density exceeds the threshold value. Our method has a 99% accuracy rate in detecting Kochia patches using multi-spectral satellite imagery with a density threshold of 40%. The semantic segmentation model trained on RGB imagery for in-field mapping has a mean intersection over union value of up to 0.8606. These results suggest pixel-level Kochia segmentation of satellite imagery can be performed more accurately if a pixel has more than 40% Kochia mix. Our study highlights the potential of using high-resolution RGB imagery and satellite data at the farm and regional levels for effective Kochia management. Detecting Kochia early and accurately can help prevent crop damage and ensure successful crop production.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".