Closed Claudius-Appel closed 6 months ago
Consider
true_identifier_pixels <- ret$identifier$pixel.idx[((ret$identifier$pixel.idx[,"x"]>1800)
& (ret$identifier$pixel.idx[,"y"]<200)
& (ret$identifier$pixel.idx[,"y"]>5)),]
pixels <- list(identifier = length(true_identifier_pixels)/2)
# for ranges where we must assume all pixels to be valid, we can just do the shortcut and choose ret$<type>$pixel.count
pixels$complete <- ret$complete$pixel.count
pixels$drought <- ret$drought$pixel.count
pixels$green <- ret$green$pixel.count
#### Evaluate: Convert to Area, display and print counts ####
duflor::convert_pixels_to_area(pixels)
duflor::plot_array_as_image_sRGB(
duflor::HSVtoRGB(
duflor::change_pixel_color_HSV(
pixel.array = pixel.array,
pixel.idx = true_identifier_pixels,
target.color = "white",
mask_extreme = T
)
)
)
I think leaving this to the user is probably better. It might be much more annoying however. I need to think about how to incorporate this into the GUI - currently, drawing a rectangle and double-clicking it is already assigned a function. How would you define pinkdot-limits?
Functionally the same as crop image
:
crop image
numericInput
s to be modified by hand, or by the rectangle-selection aboveexecute_analysis()
>execute_single()
/execute_multiple()
to filter out hits which are not part of this rectangle. Decide how to define this so it is limited to use with identifier_dot
.
When loading all images, we currently have the issue of rogue pixels being perceived as red.
You could then define an x/y range within which pink pixels are considered, and all pixels outside that range are not counted.
Maybe?