Hello,
First of all, thank you very much for providing this powerful and freely available analysis tool. I’m a master’s student currently working on a project using micaToolbox and have a few questions related to image analysis:
1. Chart Measurement
I’m using our own color checker to generate a cone catch model.
Regarding the light source, I read the forum suggestion:
“Using the same light source between spec and camera (ideally the sun on a clear day, failing that some kind of very stable broadband light source a decent distance from the surface so that each pastel receives equal irradiance).”
In our workflow, I photographed the chart outdoors under natural sunlight, but measured reflectance indoors using the spectrometer’s built-in light source (Ocean Optics DH-2000-BAL, 215–2500 nm) in a dark room. I kept a consistent distance and angle between the probe and the color checker during measurements.
To check the accuracy of our cone catch model, I compared the cone catch values generated from our reflectance data with those from your provided chart (Fig. 1). I’m unsure whether the differences between the two outputs are significant. Would these differences indicate that our reflectance measurements are unsuitable for generating a valid cone catch model and conducting image analysis?
2. Pixel Values from RNL Ranked Filter Images
In my analysis (original image shown in Fig. 2), I generated a cone catch image, applied AcuityView and the RNL Ranked Filter in QCPA framework, then measuring pixel values in the cone catch and filtered image. I noticed that some ROIs (ROI 1 and 2 in Fig. 3) had normal pixel values in the cone catch image (0~1) but showed extremely large values (up to 7 digits) in the RNL filtered image. Others (ROI 3 ~ 8 in Fig. 3) already had high values in the cone catch image, which became even larger after filtering. I’m unsure how to interpret these extremely high pixel values.
These high pixel values typically occur in bright regions of the image (see red arrows in Fig. 4 for example). I’ve confirmed that the ROIs are not overexposed, and the lighting between the standards and the ROIs is uniform.
Apologies if this has already been addressed elsewhere on the website. I truly appreciate your time and support.
Thank you very much!
Best regards,
Yung Chi