The Study Of Real-time Image Processing And Optimization For Weak And Small Target Based On DM642

Because of its low signal-to-noise ratio and background complexity,the detection of weak targets in complex background has been a technical problem in the field of image processing.And MATLAB,that is a simulating method,has more distance to engineering application in the detection of weak and small targets.Based on this,this paper depended on the National Natural Science Foundation of China ―The automatic detection method and experimental study of hydrogen content in magnesium melt based on real-time image processing‖(51374007).The real-time detection algorithm of weak target is studied by using the image processing system platform with DM642 chip as the core.The differences between the algorithms are analyzed experimentally and the real-time detection algorithms of weak and small targets and system configuration were optimized.This layered the foundation of the real-time detection of weak and small targets under complex background in the engineering application.Based on the analysis of image gray scale characteristics and combined with Prewitt operator and background prediction algorithm,an adaptive weak target detection algorithm based on DSP is proposed.In the image processing system platform with DM642 chip as the core,the algorithm using for dim and weak target detection was programmed with C language on Code Composer Studio 3.3 which is an integrated development environment.The experimental results show that the algorithm can not only detect the weak target but also improve the detection speed.Aiming at the problem of weak detection of sequence images,the algorithm of dim and weak target detection which combined two frame difference and modified semi-causal support region based on DSP was proposed,which not only successfully detects the weakness of magnesium melt Bubble targets,but also in the image noise suppression and detection speed more advantages.At the same time,according to the timeliness of DSP weak target detection,the optimization research of detection algorithm of weak and small targets was proposed from the aspects of background prediction algorithm,hardware and software structure configuration and program code.The experimental results show that the optimized detection algorithm has been greatly improved in the calculation speed.The algorithm of the paper successfully detected the small bubble target in the magnesium melt on the DM642 chip as the core of the image processing system platform,it's not only overcame the drawbacks of MATLAB software simulation but also layered the foundation of real-time,accurate and efficient detection of hydrogen content in magnesium melt with embedded chip.In addition to the optimization of the weak target detection algorithm and the system platform,it has important guiding significance for real-time detection of weak and small targets optimization research based on DSP.