How would you model this "strand" clustering problem? [P]

I'm currently building a computer vision application. I've managed to successfully train a YOLO model to detect the object I'm interested in for my videos. The image above shows some visualisations of the YOLO model outputs for some of the videos. I want to essentially cluster these strands in the image into groups based on their separation distance and return a string telling me the number of strands in each group from left to right (e.g. 1-2-3