Based On OMAP3530Stable Platform Image Processing Algorithm Design And Implementation

The guidance technology based on stable platform for image processing hasbeen widely applied to various equipments in modern war, which makes thedevelopment prospect better and better. Target detection and tracking is one of thekey technologies of image processing, and it is of academic value and practicalsignificance to research the image detection and tracking technology based onembedded system. According to the requirement of basic skills training ofprofessional application for applied graduate and combined with a research project ofTV seeker system of a missile, this paper proposes the image processing algorithmwith OMAP3530embedded development board carrying on the stable platform.Firstly, this paper briefly introduces the stable platform and its carryingOMAP3530embedded development system. The main controller of the stableplatform is TI OMAP3530embedded mini development board, which carriesARM+DSP dual-core processor. The above design greatly improves the precisionand efficiency of image processing. Moreover, the paper introduces the developmentenvironment, dual-core communication mechanism and the software developmentprocess.Moreover, this paper designs the algorithm for target detection, identificationand tracking. Considering the design requirements, the background frame differentialmethod and the finite difference method to detect the target is mainly introduced.And the algorithm based on color features for target identification is introduced.In addition, this paper focuses on the Cam shift target tracking algorithm, andproposes a method based on frame differential method and Cam shift for a targetwith deterministic initial position. After that, the paper proposes to determine theadaptive threshold using edge detection based on wavelet.Lastly, as for the OMAP3530controller, this paper applies the designedalgorithm for image identification and tracking to practical stable platform system,and achieves the design and implementation of the software. After theimplementation of image recognition and tracking algorithm in DSP, and the designof multi-threaded applications on ARM, this paper complete the dual-core datacommunication and DSP algorithm call, and verify the image detection and trackingon stable platform.