An Image Segmentation for the embedded Low Level CMOS Sensor Based on Image Enhancement

In this paper, we propose an embedded system for extracting assigned objects from an RGB image taken by the low level CMOS lens, which is installed in the embedded system. The system uses ARM9 PXA270 as the development platform and combines image preprocessing, image enhancement technology and the fuzzy calculating method to extract the assign objects. The peripheral hardware of ARM9 PXA270 are key pad, four colors LCD, low level CMOS lens, and touch screen TFT-LCD. Although the system only use a low level CMOS sensor (cost cheaper), but it can take high quality images like a high level CMOS sensor (more expensive) done. One can just use a touch pen to draw a close contour on the touch screen of the embedded system to enclose the object that one wants to extract. Then the assigned objects are going to be extracted. In order to test our system’s practical function, we have experimented with various sequences of color images. The experimental results proved that our system can quickly and efficiently extract the desired object from the RGB color