Poaching is a persistent issue that threatens many of earth’s species including therhino. The methods used by poachers are varied, but many use guns to carry outtheir illegal activities. Gunfire is extremely loud and can be heard for kilometres.This thesis investigates whether it is possible to aid anti-poaching efforts in Kenyawith a gunshot detection and estimation device using an array of microphones. Ifsuccessful, the device could be placed around the savannah or any exposed areaand warn if poaching is taking place in the nearby. If a shot is fired within theaudible range of the device’s microphones, a trained machine learning algorithmdetects the shot on the edge using a microprocessor. The detection runs in realtime and achieved an accuracy of 93% on an unbalanced data set, where themajority class was the one without gunshots. Once a detection has been made, thereceived signal power to each microphone is used to produce a direction of arrivalestimate. The estimate can produce an angle estimate with a standard deviationof 66.78° for a gunshot, and with a standard deviation of 7.65° when testing themodel with white noise. Future implementations could use several devices thatdetected the same event, and fuse their estimates to locate the shooter’s position.All of this information, as well as the sound file, can be used to alert and assistlocal wildlife services. The challenges of this project have been centred aroundmaking a system run in real time with only a microprocessor on the edge, whilealso prioritizing low cost components for future deployment. / Project Ngulia

This master’s thesis investigates the application of advanced machine learning models for the identification and classification of Swedish predators using camera trap images. With the growing threats to biodiversity, there is an urgent need for innovative and non-intrusive monitoring techniques. This study focuses on the development and evaluation of object detection models, including YOLOv5, YOLOv8, YOLOv9, and Faster R-CNN, aiming to enhance the surveillance capabilities of Swedish predatory species such as bears, wolves, lynxes, foxes, and wolverines. The research leverages a dataset from the NINA database, applying data preprocessing and augmentation techniques to ensure robust model training. The models were trained and evaluated using various dataset sizes and conditions, including day and night images. Notably, YOLOv8 and YOLOv9 underwent extended training for 300 epochs, leading to significant improvements in performance metrics. The performance of the models was evaluated using metrics such as mean Average Precision (mAP), precision, recall, and F1-score. YOLOv9, with its innovative Programmable Gradient Information (PGI) and GELAN architecture, demonstrated superior accuracy and reliability, achieving an F1-score of 0.98 on the expanded dataset. The research found that training models on images captured during both day and night jointly versus separately resulted in only minor differences in performance. However, models trained exclusively on daytime images showed slightly better performance due to more consistent and favorable lighting conditions. The study also revealed a positive correlation between the size of the training dataset and model performance, with larger datasets yielding better results across all metrics. However, the marginal gains decreased as the dataset size increased, suggesting diminishing returns. Among the species studied, foxes were the least challenging for the models to detect and identify, while wolves presented more significant challenges, likely due to their complex fur patterns and coloration blending with the background.