sikuli-driver team mailing list archive

[RaiMan <question251207@xxxxxxxxxxxxxxxxxxxxx>]

Question #251207 on Sikuli changed:
https://answers.launchpad.net/sikuli/+question/251207

    Status: Open => Answered

RaiMan proposed the following answer:
at 1*
SikuliX currently has no compound feature to find such visuals by simply ignoring the inner part. Having images for every variant and check each one after another would surely last too long.
So the only chance is to concentrate on a characteristic detail (the lower part with the arrow) and then check with other common details of the cloud (e.g. the upper edge) when found. This is a bit more complex, but can be very fast (see 2*)
Another problem you will have is the varying background behind the clouds. Since the cloud is not a rectangular area, there will always be some background in the captured image.
Since Sikuli currently does not support non-rectengular objects (or some outer parts being transparent), the only chance is, to reduce the background in captures to a minimum (10-20% of the whole image).

at 2* There is no tutorial, but faq 1607 ( and the one or other faq additionally)
The rule of thumb: the smaller the region and the more equal in size region and image are, the faster the search. The fastest achievable searches are some milliseconds regions smaller than 100 x100.
The general approach simply is, to calculate search regions relative to already found matches (which in fact are regions with some additional attributes). There are some Region functions to support this (e.g. below).

at 3* Since Sikuli workflows simulate a user on a machine, no other user or app can use screen, mouse or keyboard concurrently on that machine.
So yes, there is no other way, than to run Sikuli scripts on a different machine (virtual or real), when you want to use a machine concurrently.

-- 
You received this question notification because you are a member of
Sikuli Drivers, which is an answer contact for Sikuli.