Research On Infrared Moving Target Detection Based On TMS320C6678

Infrared moving target detection is an important research topic in air defense reconnaissance warning and intelligent video surveillance.For a long time,limited by the development level of embedded hardware,the balance between speed and precision can not be easily achieved.Taking advantage of multi-core DSP in image processing,this paper studies and realizes infrared moving target detection based on TMS320C6678.In this paper,the types and basic principles of the existing algorithms for moving target detection are summarized,then the source of the image data is briefly explained,the infrared image is preprocessed,and the main design of this paper is started.Aiming at the problem that classical algorithm GMM,ViBe and so on have poor detection effect in complex moving background and filter small targets while filtering noise during subsequent processing,the concept of structure tensor is introduced in this paper and an infrared moving target detection algorithm combined with temporal and spatial structure tensor is proposed.The software simulation results show that the proposed algorithm has good adaptability to the floating clouds,swaying trees and fluctuating water surfaces in the background,and has high detection accuracy and low missing rate.In order to further demonstrate the practical value of infrared moving target detection,the proposed algorithm is implemented in C6678 by using multi-core parallel processing method,and an infrared moving target detection platform based on C6678 is built and optimized.Running tests shows that the data flow in the platform is correct,C6678 can detect moving objects stably and accurately,and the detection speed can satisfy the visual continuity of human eyes.