Automated Bullet Dimension Extraction using Computer Vision

This project automates the precise measurement of bullet dimensions—such as overall length, case diameter, rim diameter, and ogive profile—directly from images or video frames. The pipeline performs camera calibration, perspective correction, and scale estimation using fiducials or known reference rulers. Robust segmentation and contour analysis isolate the bullet, while sub-pixel edge detection and geometric fitting (lines, circles, ellipses, spline curves) yield accurate dimension estimates. Results are overlaid on the image and exported to CSV/JSON for downstream reporting and quality control.

Built with OpenCV and numerical optimization, the system is resilient to varying lighting, backgrounds, and minor occlusions. Batch processing enables high-throughput inspection; outlier rejection and uncertainty bounds provide metrology-grade confidence. Optional modules include: automatic orientation detection, multi-view fusion for 3D-esque accuracy, and integration with a dashboard to visualize distributions and tolerances over time. The outcome is a repeatable, fast, and objective measurement process that reduces manual effort and improves consistency across production or forensic workflows.