Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Computer Vision Pipeline for Iterative Bullet Hole Tracking in Rifle Zeroing

Robert M. Belcher, Brendan C. Degryse, Leonard R. Kosta, Christopher J. Lowrance

TL;DR

The paper tackles the challenge of automating rifle zeroing by distinguishing bullet holes across successive firing iterations from line-of-fire imagery. It introduces an end-to-end CV pipeline that combines YOLOv8 for small-object bullet-hole detection, ORB-based perspective normalization for cross-view alignment, and an IoU-based method to assign holes to their corresponding iteration, supported by a novel data augmentation strategy that removes holes to simulate sequences. Key contributions include a diverse dataset from real and synthetic sources, a reverse augmentation scheme, and a robust preprocessing workflow that enhances detection and tracking performance. The results demonstrate strong per-component accuracy and practical potential for image-based, real-time feedback, with broader applicability to temporal differentiation of visually similar objects in other domains.

Abstract

Adjusting rifle sights, a process commonly called "zeroing," requires shooters to identify and differentiate bullet holes from multiple firing iterations. Traditionally, this process demands physical inspection, introducing delays due to range safety protocols and increasing the risk of human error. We present an end-to-end computer vision system for automated bullet hole detection and iteration-based tracking directly from images taken at the firing line. Our approach combines YOLOv8 for accurate small-object detection with Intersection over Union (IoU) analysis to differentiate bullet holes across sequential images. To address the scarcity of labeled sequential data, we propose a novel data augmentation technique that removes rather than adds objects to simulate realistic firing sequences. Additionally, we introduce a preprocessing pipeline that standardizes target orientation using ORB-based perspective correction, improving model accuracy. Our system achieves 97.0% mean average precision on bullet hole detection and 88.8% accuracy in assigning bullet holes to the correct firing iteration. While designed for rifle zeroing, this framework offers broader applicability in domains requiring the temporal differentiation of visually similar objects.

A Computer Vision Pipeline for Iterative Bullet Hole Tracking in Rifle Zeroing

TL;DR

The paper tackles the challenge of automating rifle zeroing by distinguishing bullet holes across successive firing iterations from line-of-fire imagery. It introduces an end-to-end CV pipeline that combines YOLOv8 for small-object bullet-hole detection, ORB-based perspective normalization for cross-view alignment, and an IoU-based method to assign holes to their corresponding iteration, supported by a novel data augmentation strategy that removes holes to simulate sequences. Key contributions include a diverse dataset from real and synthetic sources, a reverse augmentation scheme, and a robust preprocessing workflow that enhances detection and tracking performance. The results demonstrate strong per-component accuracy and practical potential for image-based, real-time feedback, with broader applicability to temporal differentiation of visually similar objects in other domains.

Abstract

Adjusting rifle sights, a process commonly called "zeroing," requires shooters to identify and differentiate bullet holes from multiple firing iterations. Traditionally, this process demands physical inspection, introducing delays due to range safety protocols and increasing the risk of human error. We present an end-to-end computer vision system for automated bullet hole detection and iteration-based tracking directly from images taken at the firing line. Our approach combines YOLOv8 for accurate small-object detection with Intersection over Union (IoU) analysis to differentiate bullet holes across sequential images. To address the scarcity of labeled sequential data, we propose a novel data augmentation technique that removes rather than adds objects to simulate realistic firing sequences. Additionally, we introduce a preprocessing pipeline that standardizes target orientation using ORB-based perspective correction, improving model accuracy. Our system achieves 97.0% mean average precision on bullet hole detection and 88.8% accuracy in assigning bullet holes to the correct firing iteration. While designed for rifle zeroing, this framework offers broader applicability in domains requiring the temporal differentiation of visually similar objects.
Paper Structure (19 sections, 1 equation, 2 figures, 1 table)

This paper contains 19 sections, 1 equation, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Bullet Hole Detection in Prior Work
  4. YOLO for Object Detection
  5. Tracking Across Images with IoU
  6. Target Segmentation via ORB
  7. Methods
  8. Dataset Construction
  9. Data Augmentation and Preprocessing
  10. Bullet Hole Detection Using YOLOv8
  11. Tracking Bullet Holes Across Firing Sequences
  12. Evaluation Metrics
  13. Results
  14. Segmentation Accuracy
  15. Bullet Hole and Target Detection
  16. ...and 4 more sections

Figures (2)

  • Figure 1: Overview of the bullet hole tracking pipeline.
  • Figure 2: Example output over two firing iterations. The top row shows images from the first firing iteration, and the bottom row shows images from the second iteration. These examples were collected indoors as part of a controlled evaluation.