i'm doing project of smart intercom system on raspberry pi camera module. in 1 script i'm doing opencv operations , displaying frames, in taking picture , sending via email... problem is, once i've accessed camera 1 script, can not again other one. in terminal i'm getting obvious message device busy: "failed enable connection: out of resources (other memory)". have of done before? i'm pretty sure it's possible, still haven't found answer 
it not possible 2 instances of camera component enabled @ same time, nor 2 processes share same component handle pointer (each process has own virtual address space, 2 don't map). not going change.
conceptually throws many questions: instance takes priority? if both try use same output stream different parameters? how choose appropriate sensor mode? how resolve differences in settings exposure mode, awb, etc?
if need 2 functions linked off same camera simultaneously need merge 2 functions same process, or write proxy service accepts multiple requests according priorities , services them. first option easier second.
don't language or tools using retrieve images beyond "a script". raspivid gained ability output yuv or rgb copy of preview buffers whilst encoding 90% of want. picamera (python) has allowed use splitter , resize provide multiple streams while, although have performance hit on gpu side if using big images.
conceptually throws many questions: instance takes priority? if both try use same output stream different parameters? how choose appropriate sensor mode? how resolve differences in settings exposure mode, awb, etc?
if need 2 functions linked off same camera simultaneously need merge 2 functions same process, or write proxy service accepts multiple requests according priorities , services them. first option easier second.
don't language or tools using retrieve images beyond "a script". raspivid gained ability output yuv or rgb copy of preview buffers whilst encoding 90% of want. picamera (python) has allowed use splitter , resize provide multiple streams while, although have performance hit on gpu side if using big images.
raspberrypi
Comments
Post a Comment