DSP System Design And Implementation Of Depth Form Defocus

The technology of DSP is being applied to much field, such as communication technology, electronic technology, computer technology, etc. There is important significance to improve the teaching of the practice of electronic and information engineering with the technology of DSP.In the study of computer vision, the measure of scene depth information mostly bases on focused image, the main method is stereoscopic vision and motion vision. Although the measuring accuracy of these methods is very high, it also makes some disadvantages that the measuring process is complex and the calculated amount is very large, so real time quality is out of condition and it is difficult to satisfy the measurement request of common motion object. The ranging principle of defocus image is based on that if the locality of the object departures from the focal plane of camera, it will form blurry defocus image in image plane, and the more the departure distance, the depth the blurry extent of image defocus, with the making use of this feature and some provisional parameters, we can calculate the range information of object. For this method belongs to monocular vision, it avoids the unefficient problems of the correspondence of stereoscopic vision feature point. Thereby in relation to stereoscopic vision, this method has practical application value in many fields.In this system design, we take the core in the design of Floating-Point DSP system board, at the same time, integrate some circuits together, such as CPLD, CMOS image sensor, step motor and control of drive, serial communications, thus it will construct a set of intact depth from defocus system. In this paper, we probe into DSP system design of depth form defocus, and state overall design process from these aspects systematically: algorithm conception, hardware design, software debugging and performance analysis. We put emphasis upon systemic hardware design, discuss the sequence control of CPLD, Register control of image sensor, and hardware exploiture, software realization, empirical analysis of DSPs, etc.