Daniel Queirós Silva

Abstract
To tackle wildfires and improve forest biomass management, cost effective and reliable mowing and pruning robots are required. However, the development of visual perception systems for forestry robotics needs to be researched and explored to achieve safe solutions. This paper presents two main contributions: an annotated dataset and a benchmark between edge-computing hardware and deep learning models. The dataset is composed by nearly 5,400 annotated images. This dataset enabled to train nine object detectors: four SSD MobileNets, one EfficientDet, three YOLO-based detectors and YOLOR. These detectors were deployed and tested on three edge-computing hardware (TPU, CPU and GPU), and evaluated in terms of detection precision and inference time. The results showed that YOLOR was the best trunk detector achieving nearly 90% F1 score and an inference average time of 13.7ms on GPU. This work will favour the development of advanced vision perception systems for robotics in forestry operations.

Abstract
The efficiency of agricultural practices depends on the timing of their execution. Environmental conditions, such as rainfall, and crop-related traits, such as plant phenology, determine the success of practices such as irrigation. Moreover, plant phenology, the seasonal timing of biological events (e.g., cotyledon emergence), is strongly influenced by genetic, environmental, and management conditions. Therefore, assessing the timing the of crops' phenological events and their spatiotemporal variability can improve decision making, allowing the thorough planning and timely execution of agricultural operations. Conventional techniques for crop phenology monitoring, such as field observations, can be prone to error, labour-intensive, and inefficient, particularly for crops with rapid growth and not very defined phenophases, such as vegetable crops. Thus, developing an accurate phenology monitoring system for vegetable crops is an important step towards sustainable practices. This paper evaluates the ability of computer vision (CV) techniques coupled with deep learning (DL) (CV_DL) as tools for the dynamic phenological classification of multiple vegetable crops at the subfield level, i.e., within the plot. Three DL models from the Single Shot Multibox Detector (SSD) architecture (SSD Inception v2, SSD MobileNet v2, and SSD ResNet 50) and one from You Only Look Once (YOLO) architecture (YOLO v4) were benchmarked through a custom dataset containing images of eight vegetable crops between emergence and harvest. The proposed benchmark includes the individual pairing of each model with the images of each crop. On average, YOLO v4 performed better than the SSD models, reaching an F1-Score of 85.5%, a mean average precision of 79.9%, and a balanced accuracy of 87.0%. In addition, YOLO v4 was tested with all available data approaching a real mixed cropping system. Hence, the same model can classify multiple vegetable crops across the growing season, allowing the accurate mapping of phenological dynamics. This study is the first to evaluate the potential of CV_DL for vegetable crops' phenological research, a pivotal step towards automating decision support systems for precision horticulture.

Abstract
The world wine sector is a multi-billion dollar industry with a wide range of economic activities. Therefore, it becomes crucial to monitor the grapevine because it allows a more accurate estimation of the yield and ensures a high-quality end product. The most common way of monitoring the grapevine is through the leaves (preventive way) since the leaves first manifest biophysical lesions. However, this does not exclude the possibility of biophysical lesions manifesting in the grape berries. Thus, this work presents three pre-trained YOLO models (YOLOv5x6, YOLOv7-E6E, and YOLOR-CSP-X) to detect and classify grape bunches as healthy or damaged by the number of berries with biophysical lesions. Two datasets were created and made publicly available with original images and manual annotations to identify the complexity between detection (bunches) and classification (healthy or damaged) tasks. The datasets use the same 10,010 images with different classes. The Grapevine Bunch Detection Dataset uses the Bunch class, and The Grapevine Bunch Condition Detection Dataset uses the OptimalBunch and DamagedBunch classes. Regarding the three models trained for grape bunches detection, they obtained promising results, highlighting YOLOv7 with 77% of mAP and 94% of the F1-score. In the case of the task of detection and identification of the state of grape bunches, the three models obtained similar results, with YOLOv5 achieving the best ones with an mAP of 72% and an F1-score of 92%.

Abstract
Selective thinning is a crucial operation to reduce forest ignitable material, to control the eucalyptus species and maximise its profitability. The selection and removal of less vigorous stems allows the remaining stems to grow healthier and without competition for water, sunlight and nutrients. This operation is traditionally performed by a human operator and is time-intensive. This work simplifies selective thinning by removing the stem selection part from the human operator's side using a computer vision algorithm. For this, two distinct datasets of eucalyptus stems (with and without foliage) were built and manually annotated, and three Deep Learning object detectors (YOLOv5, YOLOv7 and YOLOv8) were tested on real context images to perform instance segmentation. YOLOv8 was the best at this task, achieving an Average Precision of 74% and 66% on non-leafy and leafy test datasets, respectively. A computer vision algorithm for automatic stem selection was developed based on the YOLOv8 segmentation output. The algorithm managed to get a Precision above 97% and a 81% Recall. The findings of this work can have a positive impact in future developments for automatising selective thinning in forested contexts.

Abstract
Object identification, such as tree trunk detection, is fundamental for forest robotics. Intelligent vision systems are of paramount importance in order to improve robotic perception, thus enhancing the autonomy of forest robots. To that purpose, this paper presents three contributions: an open dataset of 5325 annotated forest images; a tree trunk detection Edge AI benchmark between 13 deep learning models evaluated on four edge-devices (CPU, TPU, GPU and VPU); and a tree trunk mapping experiment using an OAK-D as a sensing device. The results showed that YOLOR was the most reliable trunk detector, achieving a maximum F1 score around 90% while maintaining high scores for different confidence levels; in terms of inference time, YOLOv4 Tiny was the fastest model, attaining 1.93 ms on the GPU. YOLOv7 Tiny presented the best trade-off between detection accuracy and speed, with average inference times under 4 ms on the GPU considering different input resolutions and at the same time achieving an F1 score similar to YOLOR. This work will enable the development of advanced artificial vision systems for robotics in forestry monitoring operations.