Design And Implementation Of High Resolution InSAR Real-Time Processing

The interferometric synthetic aperture radar(InSAR)is a branch developed from the synthetic aperture radar(SAR).InSAR uses a couple of antennas to observe the same scene of the grond to obtain two SAR images with coherence,then it gets the digital elevation map by processing these SAR images.The working condition of InSAR is not limited by the weather and time.InSAR plays a very important role in the military and mapping fields.This paper studies the real-time processing of high-resolution interferometric synthetic aperture radar.As the resolution of InSAR is getting higher and higher,the amount of data that the processing platform needs to process is also increasing.This leads to an increase in hardware design and software processing complexity,and the single-chip DSP cannot meet the real-time requirements.Therefore,this paper designs and implements a high-resolution InSAR real-time processing flow on a real-time processing platform based on two TMS320C6678 interconnects.The main content of this article can be summarized as follows: 1.This paper studies the data processing module of multi-core DSP TMS320C6678.The advantages and disadvantages of multi-core DSP based on the main and auxiliary models and the data flow model are discussed.According to the characteristics of serial processing of interferometric synthetic aperture radar,the data flow model is selected.Then,the problem of load balancing between cores in the data flow model is further explored.2.According to the interconnection characteristics of multi-chip DSP,this paper presents the data interaction and task coordination method of multi-chip DSP in high-resolution InSAR real-time signal processing design.One DSP can not meet the real-time requirements of the system due to the high resolution and large amount of data.It requires two or more DSPs to work together and process the same image in parallel to reduce the processing time and meet the real-time processing requirements.This paper analyzes the data reading and writing efficiency of Hyperlink in detail through experiments and draws conclusions: when data migration is required frequently,the performance of the EDMA3 write operation initiated by the local device is better than the performance based on direct CPU access.This provides an experimental basis for solving the problem of large data transmission and task coordination when two DSPs process high resolution InSAR.3.This paper studies the data interaction process when two DSPs cooperate in InSAR processing.In order to ensure the continuity of the final results of high-resolution InSAR data in large scenes,according to the characteristics of SAR image segmentation,interference phase unwrapping,absolute phase determination and fuzzy number transfer,the paper gives the optimal strategy of two pieces of DSP information transmission and data interaction in the whole process data processing.4.Aiming at the phase unwrapping problem under the condition of high-resolution InSAR and large data volume,this paper proposes a phase unwrapping method based on splitting data into squares and "class-like regional growth method".The simple processing method of splitting by range or azimuth is unable to meet the precision stability requirements.In order to solve the contradiction between large data and small memory device,this paper proposes a method of splitting wrap phase into squares,and discusses the splitting principle in tetial.This paper designs and implements the core load optimization scheme,which effectively solves the problem of time consumption in the conventional load balancing scheme.The "class-like regional growth method" is proposed for splicing the squares into a large image.The processing results of the airborne measured data verify the effectiveness of the proposed method and design.