使用渲染脚本进行图像处理

Question

I am trying to have a renderscript calculate the min and max of a photos pixels rgb. I don't need the min and max in my java if I can keep the memory in the RS. How will i know if i have the memory to keep the min and max on the gpu? I am snapping photos every 3 seconds and i want to know if some thing is moving in the cameras field of view. I like the idea of learning the scene with the min and max values of the color and comparing the next snapshot with the min and max. I am familiar with most examples on the web relating to renderscripts but I don't know the best way to handle this. A code snippet would be awesome.

Thanks

Answer 1

Renderscript for gpu is supported only on a handful devices right now. So depending on the gpu might not be what you're looking for. Also, renderscript is only useful to speed up the tight loops in your code (like looping over an entire image).

From what I can think, you need to move your renderscript data into your program to do any kind of decision making.

You can make an allocation with just two numbers (bytes, floats, etc). Pass it to the renderscript and let it iterate over the image. Then the renderscript populates allocation and you can access it inside Java. You can maintain an array in your Java code

PS: Using an allocation is important because you can't access individual variables from a renderscript in your Java code (apparently this was a design choice). So something like this would be inaccessible to your Java code:

int max, min     // in renderscript