Research platform for professional video processing

Research platform for professional video processing

Simplicity is an excellent design concept

The MIMO DSP Platform is designed with a simple and clear interface that minimizes the learning cost of development, and is used as a research platform for professional video processing by companies involved in professional observation systems.

The minimum implementation required for computational processing is shown below.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40


class Channel1: public mimodsp::plugin::IBrain
{
public:
    Channel1() = default;
    const char *Name() const override
    {
        return "Channel1";
    }
    bool open(const std::string &setup) override
    {
        handle = quadcamera_open();
        return true;
    }
    bool process(
        const mimodsp::plugin::VideoFrameRack &input,
        mimodsp::plugin::VideoFrameRack &output) override
    {
      if ((input.getChannels() >= 4) && (output.getChannels() >= 1)) {
        /*
         * Execute the GPU kernel.
         */
        quadcamera_execute_uyvy_to_uyvy(handle,
            vf1.getPictureWidth(),
            vf1.getPictureHeight(),
            input.getChannelVideoFrame(0).getBuffer(),
            input.getChannelVideoFrame(1).getBuffer(),
            input.getChannelVideoFrame(2).getBuffer(),
            input.getChannelVideoFrame(3).getBuffer(),
            output.getChannelVideoFrame(0).getBuffer());
      }
      return true;
    }
    bool close() override
    {
        quadcamera_close(handle);
        return true;
    }
private:
    QUADCAMERA *handle;
};

Video input/output processing can be done with built-in plug-ins, so with just a few dozen lines of code, you can realize four channels of full HD 60p input and real-time image processing with the GPU. Of course, expansion of the number of channels can be handled with a few implementation changes.