memory phenomenon

Hi. I made a m-code for image processing, but I detect some strange effect. For example I use the following code:

vid = videoinput('winvideo', 1,'YUY2_320x240');
set(vid,'ReturnedColorSpace','rgb','TriggerRepeat',1000,'FramesPerTrigger',1);
vid_src=getselectedsource(vid);
set(vid_src,'BacklightCompensation','off','ExposureMode','manual','WhiteBalanceMode','manual','WhiteBalance',6500);
triggerconfig(vid,'manual');
start(vid);
for step=1:1:inf
tic
trigger(vid);
A=getdata(vid,1);
B=rgb2hsv(A);
.
.
. %some code
.
.
figure(4);imshow(A);hold on;
plot(q,p,'Marker','*','Color','r');
flushdata(vid,'triggers');
toc
end

I observe the result for the time of execution and I saw that every cycle the time is increasing and become bigger and bigger. So algorithm is working slower and slower. I can't understand why this is happening. I use a flushdata to clear the memory, but it dosen't matter whether I use it or not. When the data are logged in the memory GETDATA function clear it when I take the data into the workspace. Can somebody tell me what is the reason and how I can fix the problem?

Answer 1

[Walter Roberson]

You do not show figure 4 as having its contents cleared, so the "hold on" would result in another graph being added to the figure for each loop iteration.

[Hristo Alexiev]

Yes, but this is not the problem. The problem is the time of execution of every next iteration. I can't understand what's happening in the memory where I log the frames, use them, and then I delete them.

[Walter Roberson]

You only show one loop in your code, and that loop does not delete the plots you create with "hold on" set. The *plots* accumulate in memory and slow everything down.

[Jan]

As Walter said already: Are you sure, that you delete the objects correctly?

[Hristo Alexiev]

Thank You for your help. When I use "hold off" to clear the accumulated plots the program start working faster and don't slowing at the time of working.

Best Regards

Hristo