Image processing for low-power microcontroller application

Abstract

This paper focuses on the signal processing of the image signal from the camera system, which is applied to the low-power microcontroller. Image signal loading and processing is realized on the low-power Freescale DSP56F805, which is combination of the digital signal processor and the controller. There are described data and control signals connected from the digital camera and connection to specified pins of the digital signal processor. The number of scanning images per second depends on the clock signal frequency that is generated inside digital signal processor. The defined synchronization clock ensures correctly data capturing form the camera. An external RAM memory is necessary to use because of the insufficient size of the storage space for scanned images. There is implemented algorithm of frequency analyses for verification of low-power microcontroller capability. The amplitude and phase frequency spectrum of the scanned image signal is obtained by 2D FFT (Fast Fourier Transform) application. The amplitude frequency spectrum contains information about colors and the phase frequency spectrum contains information about edges and transitions in the analyzed image. Key parameters and elements of the image can be analyzed by frequency spectra results. Mainly it is important for the classification and sorting of image data. The programming implementation is realized in CodeWarrior programming environment with the ProcessorExpert tool using, which are supplied by Freescale Company.