I automated the process of counting the number of bucket teeth on a loader to ensure none were missing. I utilized a few-shot object detection technique known as template matching. After collecting 5-10 sample images of the loader bucket from a video, I applied OpenCV’s template matching algorithm, which generated bounding boxes for the bucket with varying confidence levels. These were thresholded to minimize noise. Then, using a color histogram, I counted the number of teeth in each extracted image.