[Finding_Lanes] Fix ROI hardcoding, color space mismatch, line slope filtering, and add CLI & unit tests

[Prayas340]

Description

The Finding_Lanes project contains several bugs, hardcoded dimensions, and missing features:

1. Hardcoded ROI Polygon Mask: roi() uses static coordinates [(200, height), (1100, height), (550, 250)], which breaks on any image or video of different resolutions (e.g. 1080p, 480p).
2. Color Space Mismatch: canny() uses cv2.COLOR_RGB2GRAY on OpenCV frames loaded in BGR format, which violates OpenCV conventions.
3. Noisy Slope Filtering & Divide-by-Zero Risk: average_lines_intercept() fails to filter out near-horizontal road artifacts and shadows (slopes near 0), and make_coordinate() lacks zero/NaN/inf safeguards.
4. Hardcoded Input Paths & Lack of CLI: Only runs on a fixed video.mp4 file; no support for CLI arguments (--image, --video, --camera, --output, --no-show).
5. Incomplete Subplot Utility (sub.py): sub.py was an orphaned placeholder script displaying the raw image without running lane detection.
6. No Unit Tests & Incomplete Documentation: No test suite exists to verify pipeline robustness, and README.md had placeholder text.

Type of issue

• Bug
• Feature (New Capabilities / Unit Tests)
• Documentation

Checklist:

• I have read the project guidelines.
• I have checked previous issues to avoid duplicates.
• This issue will be meaningful for the project.