There are two parts of the system design based on the python language:

1. The image data extraction and processing

  Firstly, we installed the openCV library in the R-pi and utilize the default data structure and functions of openCV to save and process the image data taken by the camera module. Considering the screen resolution (320*240) and image quality, we down sample the size to 640*480 for restoring and 320*240 for showing on the screen.

  In addition, we can create some special effects on the saved images, e.g., to change the pixel value of RGB color space and produce special effects on the images. Furthermore, the more advanced applications, such as face recognition and motion detection in surveillance system can also be realized. For image processing, we support two types of photo style: One is vintage style, and the other one is pencil sketch style. For vintage processing, we perform operations on RGB color space. We calculate the new RGB values for each pixel by weighted sum of the original RGB values. The equations are listed as follows:

newRed = (inputRed * .393) + (inputGreen *.769) + (inputBlue * .189)

newGreen = (inputRed * .349) + (inputGreen *.686) + (inputBlue * .168)

newBlue = (inputRed * .272) + (inputGreen *.534) + (inputBlue * .131)

(Referenced from [7])

For Mosiac style, we operate only on grayscale image. First we separate the image into several blocks and compute the intensity histogram within its own block. Then we rescale the range for each block to 0~255.

2. The interface of operation system

  We introduce pygame library to help build our operation platform. The image shown on the screen could be viewed as a “screen surface” where we can attach all images we want on it. For example:

      a. The image data extracted by openCV can be attached on the screen surface and keep updated. Thus we can see the real time image taken by camera.

      b. The storing images and images with post-processing can be attached on the screen surface.

     c. The control buttons can be drawn and attached on the screen surface. Besides, the alpha value (transparency level) is tunable, so the buttons will not shelter the background image.

  The mode tags are designed for users to switch to different modes. There are totally 7 modes in our system and the detail functions will be covered in the testing section. When users touch the buttons on the screen, the pygame event would be triggered and system will check the touching location to decide which function should be called. For instance, the delete, taking picture commands can be performed or the mode tag can be changed and then system jump to another operation mode.