YOLOv5 Model Application in Real-Time Robotic Eggplant Harvesting
Bibliographic record
Abstract
Deep learning studies in agricultural automation have accelerated in recent years due to its benefits such as increasing product efficiency and reducing labor force. Deep learning is a powerful tool for automation in agriculture with applications ranging from disease identification and crop yield detection to fruit ripeness classification. It helps to automate various processes in agriculture and to perform time-consuming tasks in a shorter time. It quickly processes the data required for robotic harvesting systems and makes it available to the system. In this study, a machine learning study was carried out to be used in the robotic harvesting of eggplant fruit, which is a product that can take time to select and collect in the agricultural area where it is cultivated. YOLOv5 (nano-small-medium and large models) was used for the deep learning method. All training and test metric values of the models were analyzed. It was determined that the most successful model was the model trained with YOLOv5m algorithm on images of 640 × 640 size with 12 Batches and 110 Epochs. The results of the model values were analyzed as “metrics/precision”, “metrics/recall”, “metrics/mAP_0.5” and “metrics/mAP_0.5:0.95”. These are key metrics that measure the detection success of a model and indicate the performance of the relevant model on the verification dataset. It was determined that the metric data of the “YOLOv5 medium” model was higher compared to other models. The YOLOv5m model gave the highest score with F1 score of 85.66%, precision of 95.65%, recall of 96.15%, and mAP at 0.5:0.65 of 78.80%. Hence, it was understood that “Model 3” was the best detection model to be used in robotic eggplant harvesting to separate the eggplant from branch.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
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".