Multi-Camera Calibration Expert for Basketball Court Setup
Description
Looking for an engineer with expertise in computer vision and camera calibration to design and implement a robust calibration workflow for recorded basketball games. The setup involves calibrating multiple cameras using intrinsic parameters (precomputed) and extrinsic parameters. The goal is to ensure high accuracy and consistency across various court sizes without using pre-measured points.
Requirements:
1. **Intrinsic Calibration**:
– Experience with OpenCV’s `calibrateCamera()` or `fisheye::calibrate()`.
– Familiarity with checkerboard/Charuco pattern calibration.
2. **Extrinsic Calibration**:
– Knowledge of 3D-2D correspondences using real-world references.
– Proficiency in rim and line detection via image processing (e.g., Hough transform, edge detection).
3. **Image Processing**:
– Expertise in color space manipulation (e.g., HSV/LAB).
– Use of adaptive histogram equalization for lighting normalization.
4. **Optimization Techniques**:
– Proficiency in using `cv::solvePnP` with RANSAC for pose estimation.
– Familiarity with bundle adjustment for refining global consistency.
5. **Multi-Camera Integration**:
– Experience in matching cross-view points for unified coordinate systems.
– Knowledge of multi-camera pose optimization.
6. **Validation and Metrics**:
– Ability to evaluate and minimize reprojection errors.
– Ensure stability of extrinsic parameters across frames.
7. **Robustness**:
– Capable of handling varying lighting conditions and non-standard courts.
– Design a system that works without specialized calibration objects on-site.
Nice to Have:
– Familiarity with sports analytics and real-world deployment of calibration systems.
– Experience with Python or C++ for OpenCV-based development.
—
This position is ideal for a computer vision engineer with experience in tasks requiring multi-camera calibration.
