@Bob_Schor さんは書きました:
One of the first things to do is to (as clearly as possible) describe the task. I'm not entirely sure what you want to do, nor how you want to do it.
The "known" factor is a 256 x 256 array of (presumably uniformly-spaced) "destinations" that you want to "visit" and either give a "value" to or to "get" a value from. The basic question appear to be how you traverse this 256 x 256 array.
So here's a question -- are you able to read (or write) all 65536 values (256 x 256) all at once, or do you need to "traverse" this space, one element at a time? If you are "traversing", do you go from one element to an adjacent element, moving at a uniform rate from element to (adjacent) element (and hence a fixed delta-t) or can you go from any element to any other element in a single time step? (Note that if the time step is short enough compared, say, to the relatively slow time of human vision, which lets us watch TV or a movie and not notice that the images are changing every 10 ms or so -- we don't notice the screen being "painted").
There are several ways to traverse a 256 x 256 array, depending on the timing you want to use. If you are "painting with an electron beam" (like a TV), positioning the beam is very fast (provided you can turn it on and off so you don't notice the sweeps), so you "paint" row 0, then row 1, ... then row 255, then row 0 again.
If, on the other hand, you have a stationary "beam" and are moving the "object being scanned", then you might want to move one step at a time (at a constant rate). How do you want to do this?
One way (for a single 256 x 256 scan) is the following:
- Start at Row 0, Column 0.
- Scan even Rows one column at a time, incrementing the column until you reach Column 255.
- When you finish 0, 255, increment the row (but leave the column unchanged).
- Scan odd Rows one column at a time, decrementing the column until you reach Column 0.
- When you finish (odd row), 0,
- If the Row is 255, you are done, exit.
- Otherwise, increment the row (to make it even) and go back to Step 2.
But what if you want to do multiple scans? How do you get from (255, 0) back to (0, 0)? There are many ways to do this (one is to simply do the whole algorithm again, but reverse the Y increment, i.e. basically reverse your steps -- the draw back, of course, is if the image is changing, the time-between-samples becomes a complicated function of pixel location -- my head hurts even thinking about this!
Anyway, once you clearly describe the algorithm you want to use, coding it should be a piece of cake. Here's how to do the "TV Algorithm":
- For each frame:
- Move to 0, 0
- For each of 256 rows:
- Process Pixel.(x, y)
- If x < 255, increment x and move to (x, y)
- otherwise if y < 255, increment y and move to (0, y)
- otherwise go back to Step 1 and "do it all over again"
Because moving electrons is so much faster than moving larger objects, cameras are easier to use (and program) than microscope stages.
Bob Schor
I apologize for the confusing explanation.
I want to traverse a 256 x 256 space one element at a time. We intend to get data at each location. We intend to move from adjacent element to element at a uniform speed (about 300μs).
When it reaches (x,y)=(0,255), it moves 255 in the x-axis direction and comes back 255. Next, it moves to (0,254) and continues to move in the x-axis direction as before, but now we want it to move 255 again from (0,255) in the x-axis direction and back before it reaches (0,254). Next, we want to move to (0,254) and continue the same behavior as before.
Image
Now: (x,y)=(2,254)(1,254)(0,254)(0,255)(1,255)(2,255)... (255,255)(254,255)... (0,255)(0,254)(1,254)(2,254)...
After Improvement：(x,y)=(2,254)(1,254)(0,254)(0,255)(1,255)(2,255)...(255,255)(254,255)...(0,255)(0,255)(1,255)(2,255)...(255,255)(254,255)...(0,255)(0,254)(1,254)(2,254)...
Thank you for your kind reply.
