Computer Vision Developer for Textile Analysis

I need a solution that detects fine-grained grid intersection points on scanned images of woven mesh textile, then maps a custom design onto those real intersection coordinates and outputs a print-ready file. Context: the input is a flatbed scan of real fabric with visible thread intersections — not a printed calibration pattern — so it's lower contrast and less uniform than a typical camera-calibration target. This is conceptually similar to OpenCV checkerboard/circle-grid corner detection, adapted to messier, real-world material, so it'll need genuine tuning rather than an out-of-the-box function call. Full project scope (one contract, structured in two milestones): Milestone 1 — Detection: Given sample scanned canvas images (attached), detect every thread intersection across the full image and output coordinates to sub-pixel accuracy, indexed by row/column if possible. Deliverable: a working script plus a visual overlay showing detected points on the sample scans, so accuracy can be verified by eye. Milestone 2 — Mapping and output: Using the coordinates from Milestone 1, map a design file (a grid of colors, one per intersection) onto the real detected positions, correcting for any rotation, skew, or uneven spacing in the scan. Output a print-ready image file sized and positioned to match that specific canvas. Milestone 1 is the harder, more uncertain part of this project and will be paid and approved separately before work begins on Milestone 2. What I'm looking for: Real experience with Python, OpenCV, and image processing — ideally camera calibration, corner detection, or grid/pattern recognition specifically, not general ML/AI Comfortable working with real-world, imperfect image data rather than clean synthetic input Please share an example of past work involving corner, grid, or pattern detection if you have one, and note whether you'd want to structure this as two milestones within one contract, as described above