Grid layout image analysis workflows

[HaleySchuhl]

Is your feature request related to a problem? Please describe.
Tracking seeds with their ID information and analyzing multiple objects when they are in a grid layout

Describe the solution you'd like
A handful of functions that work with upstream workflow steps like pcv.roi.* function and would hopefully enable users to analyze multiple biologically relevant units within a single image. Would hopefully work with all analyses functions, or maybe would have separate functions? Something where we could also use metadata from a CSV (info about genotype, plant ID, treatment, whatever) and would also be accessible through bar code reading ideally.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

• Identify the centers of a grid automatically. This would work downstream with pcv.roi.multi, since we can feed it a list of centers for ROIs to get drawn. This might be more useful than a labeled mask, since we want to assume the mask fed into the grid recognizing/ auto layout will not be the same as the an analysis mask.

Additional context

• Katie's maize seed analysis and linking to downstream traits collected from mature plants.
• Emma and Kellogg lab seed scattering project.
• Be able to crop and save out (pro-processing workflow) but ALSO be able to analyze individual objects within the same workflow.
• Miller lab's data

[HaleySchuhl]

First steps:

• Detecting the grid layout without as many settings. e.g. pcv.roi.multi and cluster_contours requires you to know a lot about the grid layout. Provide mask of objects, and just number of rows and columns --> be able to detect coordinates of the centers of all objects, and thus could calculate spacing in x and y direction and the starting coordinate. (Similar to pcv.transform.find_color_card? )
• Maize kernels are being individually tracked and have identifiers that will hopefully stick with them throughout their life span. We have the ability to customize labels already, but want to enable the ease with which a researcher can give other identifiers.

Next steps:

• label objects