The AMB82 microprocessor is capable of running YOLO tiny for object detection. By running the model locally instead of on the cloud, it saves time and keeps data more secure. When we found out that the senior student’s training of the original YOLOv4 tiny model was not good enough, we decided to improve it. We used a method with overlapping bounding boxes to better preserve the information of the detected objects. This improvement boosted the model's accuracy, with the mAP@0.5 score increasing from 78.95% to an excellent 90.45%. Use Case 1: Bird Watching for Ecology next to our school is a beautiful lake where people can usually only stand on the shore to watch birds. We used an eco-friendly boat equipped with the AMB82 chip to sail onto the lake. When the system detects a bird, it automatically starts recording an mp4 video of the bird’s life and sounds. This allows us to observe the birds from a closer and different perspective. This recording effort contributes to the SDGs (Sustainable Development Goals) Target 15.5 by monitoring and taking action to prevent the decline of biodiversity. Use Case 2: Protecting Farm Animals We saw a news report about stray dogs breaking into a fenced chicken coop, which resulted in the loss of hundreds of chickens and a huge financial loss for the farmer. To prevent this, we used the AMB82 to monitor the area. When the system identifies a dog trespassing, it immediately sends an email notification to the owner. This allows the farmer to arrive at the chicken coop in time to save the chickens and reduce losses. This aligns with SDGs Target 2.3 by helping to increase the productivity and income of small-scale farmers.