The project is to reliably detect defects in a sugar pack. Defects can occur on the top and sides of the pack so all visible areas need to be analyzed. The deliverable is a Matlab algorithm (with a detailed explanation and comments) that works on all attached samples. The legend for the samples 2 and 15 is provided in samples1.pdf. The same legend logic applies to all samples. All attached samples are in the jpg format; bmps are available upon request.
1. Defects in both samples must be detected.
2. The thresholding must include the top and sides, as the defects can appear anywhere.
3. The sides of the bag cannot alarm just because they are less illuminated.
4. The same settings must be applied to all samples.
5. Cracks and Artifacts A cannot false alarm. However, It might be acceptable if only the most severe cracks alarm - we need to know the limits of the analysis
6. Alarms on Artifacts B, bag ID and Lens Reflection are acceptable as these items either do not exist in real life or we have a way of removing them.
15 freelancers are bidding on average $729 for this job
Hi, I am a PhD student in the field of Computer Vision(Machine Machine and Image Processing). I am interested as this is my area of specialization. I am an expert with matlab. Please check my private message.
Hi Sir, I did something similar on wafer (semiconductor) images analysys and in rice analysis. I think I can do it in the same way. Do not hesiatte to ask me for any question. Best regards
Most image processing algorithms does not work for X-Ray images, I have many years experience in medial image processing and your problem is challenging but certainly doable! PM me for more details
Custom software development ([url removed, login to view] Administrator removed this message for containing contact details which breaches our Terms of Service)