Save Your Output to a Folder

Save Your Output to a Folder

You can specify a path to a folder to save your algorithm results as image files.

int main(int argc, const char * argv[])
{
    
    auto processFrame = [](size_t frameN, Mat &input, Mat &output){
        GaussianBlur(input, output, Size(7,7), 1.5, 1.5);
        Canny(output, output, 0, 30, 3);
    };
    
    Ptr<Input> input = new CameraInput(0);
    Ptr<Output> output = new ImageOutput("path/to/folder/");

    Processor processor;
    processor.setInput(input);
    processor.setOutput(output);
    processor.setProcess(processFrame);
    processor.run();
    
    return 0;
}

The ImageOutput class takes a string as its first argument, the folder path – where your images will be stored (e.g. "path/to/folder/").

The setOutput method sets the Processor class to use the output instance (i.e. ImageOutput).

processor.setOutput(output);

The Processor class will create a new thread to process the writing process to disk using the specified Output instance (i.e. ImageOutput). Each output image created by the lambda function will be directed to the ImageOutput instance and written to disk.

You can force the image size of your output by passing a second argument specifying the desired resolution. If you would like to use the original resolution a value of Size(-1,-1) should be passed (i.e. default value Size(-1,-1)).

Ptr<Output> output = new ImageOutput("path/to/folder/", Size(640,480));

A third argument specifies the number of digits your image file will contain (e.g. "00000.jpg" has 5 digits -default value-)

Ptr<Output> output = new ImageOutput("video.avi", Size(640,480), 5);

A fourth argument specifies the OpenCV color space conversion code to apply to the output images generated in the lambda function. By default no code conversion is applied to the images (i.e. default value - 1).

Ptr<Output> output = new ImageOutput("video.avi", Size(640,480), 5, CV_RGB2GRAY);

This is useful in the case you want to generate in a different color spectrum from the one you outputted in your lambda function. (e.g. CV_GRAY2RGB, CV_RGB2HLS, ... etc)