Browsing by Subject "YOLO"

Now showing 1 - 2 of 2

Assessing the capability of YOLO- and transformer-based object detectors for real-time weed detection
(2025) Allmendinger, Alicia; Saltık, Ahmet Oğuz; Peteinatos, Gerassimos G.; Stein, Anthony; Gerhards, Roland
Spot spraying represents an efficient and sustainable method for reducing herbicide use in agriculture. Reliable differentiation between crops and weeds, including species-level classification, is essential for real-time application. This study compares state-of-the-art object detection models-YOLOv8, YOLOv9, YOLOv10, and RT-DETR-using 5611 images from 16 plant species. Two datasets were created, dataset 1 with training all 16 species individually and dataset 2 with grouping weeds into monocotyledonous weeds, dicotyledonous weeds, and three chosen crops. Results indicate that all models perform similarly, but YOLOv9s and YOLOv9e, exhibit strong recall (66.58 % and 72.36 %) and mAP50 (73.52 % and 79.86 %), and mAP50-95 (43.82 % and 47.00 %) in dataset 2. RT-DETR-l, excels in precision reaching 82.44 % (dataset 1) and 81.46 % (dataset 2) making it ideal for minimizing false positives. In dataset 2, YOLOv9c attains a precision of 84.76% for dicots and 78.22% recall for Zea mays L.. Inference times highlight smaller YOLO models (YOLOv8n, YOLOv9t, and YOLOv10n) as the fastest, reaching 7.64 ms (dataset 1) on an NVIDIA GeForce RTX 4090 GPU, with CPU inference times increasing significantly. These findings emphasize the trade-off between model size, accuracy, and hardware suitability for real-time agricultural applications.
A novel dataset of annotated oyster mushroom images with environmental context for machine learning applications
(2024) Duman, Sonay; Elewi, Abdullah; Hajhamed, Abdulsalam; Khankan, Rasheed; Souag, Amina; Ahmed, Asma
State-of-the-art technologies such as computer vision and machine learning, are revolutionizing the smart mushroom industry by addressing diverse challenges in yield prediction, growth analysis, mushroom classification, disease and deformation detection, and digital twinning. However, mushrooms have long presented a challenge to automated systems due to their varied sizes, shapes, and surface characteristics, limiting the effectiveness of technologies aimed at mushroom classification and growth analysis. Clean and well-labelled datasets are therefore a cornerstone for developing efficient machine-learning models. Bridging this gap in oyster mushroom cultivation, we present a novel dataset comprising 555 high-quality camera raw images, from which approximately 16.000 manually annotated images were extracted. These images capture mushrooms in various shapes, maturity stages, and conditions, photographed in a greenhouse using two cameras for comprehensive coverage. Alongside the images, we recorded key environmental parameters within the mushroom greenhouse, such as temperature, relative humidity, moisture, and air quality, for a holistic analysis. This dataset is unique in providing both visual and environmental time-point data, organized into four storage folders: “Raw Images”; “Mushroom Labelled Images and Annotation Files”; “Maturity Labelled Images and Annotation Files”; and “Sensor Data”, which includes time-stamped sensor readings in Excel files. This dataset can enable researchers to develop high-quality prediction and classification machine learning models for the intelligent cultivation of oyster mushrooms. Beyond mushroom cultivation, this dataset also has the potential to be utilized in the fields of computer vision, artificial intelligence, robotics, precision agriculture, and fungal studies in general.