Real-time 3D Image Reconstruction of a 24×24 Row-Column Addressing Array: from Raw Data to Image

Chunyu Li,Jiali Yang,Xu Li,Xiaoli Zhong,Junjie Song,Mingyue Ding,Ming Yuchi
DOI: https://doi.org/10.1117/12.2216069
2016-01-01
Abstract:This paper presents a work of real-time 3-D image reconstruction for a 7.5-MHz, 24×24 row-column addressing array transducer. The transducer works with a predesigned transmit/receive module. After the raw data are captured by the NI PXIe data acquisition (DAQ) module, the following processing procedures are performed: delay and sum (DAS), base-line calibration, envelope detection, logarithm compression, down-sampling, gray scale mapping and 3-D display. These procedures are optimized for obtaining real-time 3-D images. Fixed-point focusing scheme is applied in delay and sum (DAS) to obtain line data from channel data. Zero-phase high-pass filter is used to calibrate the base-line shift of echo. The classical Hilbert transformation is adopted to detect the envelopes of echo. Logarithm compression is implemented to enlarge the weak signals and narrow the gap from the strong ones. Down-sampling reduces the amount of data to improve the processing speed. Linear gray scale mapping is introduced that the weakest signal is mapped to 0 and the strongest signal 255. The real-time 3-D images are displayed with multi-planar mode, which shows three orthogonal sections (vertical section, coronal section, transverse section). A trigger signal is sent from the transmit/receive module to the DAQ module at the start of each volume data generation to ensure synchronization between these two modules. All procedures, include data acquisition (DAQ), signal processing and image display, are programmed on the platform of LabVIEW. 675MB raw echo data are acquired in one minute to generate 24×24×48, 27fps 3-D images. The experiment on the strong reflection object (aluminum slice) shows the feasibility of the whole process from raw data to real-time 3-D images.
What problem does this paper attempt to address?