Detect 1)sheets 2)stains 3)holes from samples, identify using images, 2 & 3 in new samples with small computer + cam (C++/Python)
Relevant Skills and Experience
1) Developed solutions for real-time footpath identification from monocular cam and small computer <1.6Ghz, 2GB RAM, C++ & OpenCV.
2) Developed system for server-client model to process huge 4D point-clouds on server, and visualize on tablet or smartphone using web-browser (remote)
3) 5+ yrs with Matlab, Python, C++ coding for research papers, uni assgnmts
Regarding your concerns:
Yes, it is possible to devise such solution at low cost (hardware + software).
DATA: To begin with we need 10 videos (let's say) of ironing process (like your link, with some adjustments THAT I'LL ASK) of 10 mins each. A recording with just 30 FPS is suitable. All sheets or stains NOT needed (as are WHITE), since image processing can help much to detect rust (color, texture) that differs from cloth. Similarly, for holes (black background (ADJUSTMENTS etc.).
Adding DATA? Yes, later new sheets, stains can be identified
SetUp: 2 cams, black test-bed, software installation
Proposed Milestones
$700 AUD - 1st: Prepare data, model selection (SVM etc.), train
$500 AUD - 2nd: Test on new data, tune model
$500 AUD - 3rd: Attach GUI, clean & package code, assist in installation & run
More?
Initiate chat for discussion as I've lot more to say but the word limit obstructs it.
About me:
Research assistant at computer vision dept. AI applications, NOTE: Funds from this project shall go for paper publication, shall promote your firm, if you sponsor RnD developing tool/software
Thank you!