This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
Answering my own question, as I've learned to re-model my thinking around a better understanding of what openCv does, since. Short answer: no, one cannot interact with/control a kernel.
The challenge that motivated this question was treating an image with varying lighting, contrast and detail: the only ready-made function that does this, to my (still-limited) knowledge, is the 'adaptiveThreshold' function.
So if one has only an image with varying contrast, lighting, etc., to work with, and the analysis process is complex (beyond threshold finding: smoothing, filling gaps, finding lines and/or edges, etc.), one must first isolate that part of the image, then work on it (and paste the result to a 'processed' output image): this is basically akin to 'controlling the kernel'.
