An Experimental Study on Holographic Subsurface Radar System and Its Application to Bone Fracture Imaging

Abstract

To image the inside of the human body without surgery, X-ray, computer tomography (CT), magnetic resonance imaging, and ultrasound are frequently used. Among them, X-ray and CT are most common methods for imaging the human bones. However, they are based on ionizing radiation and may cause damage to the cells in human body. Thus, in this thesis, a bone imaging method using a holographic subsurface radar system was proposed. This method is based on non-ionizing electromagnetic waves, which is relatively safer to human. Holographic subsurface radar system produces an image according to the permittivity difference of the human tissue such as skin, fat, and bone. For a feasibility study, the body near bone was modeled as a simple one-dimensional (1D) structure, which consists of three layers; skin, fat, and bone. Each layer is associated with the frequency-dependent permittivity values. One-dimensional numerical analysis based on the plane wave theory was performed to the model, and the reflected wave intensity generated by an incident plane wave was calculated. It was shown that the largest difference between the ‘bone-less’ and ‘bone-present’ cases appears near frequency of 15 GHz. A series of experiments for three-dimensional (3D) structure was performed using a holographic subsurface radar system. The system consists of an open-ended rectangular waveguide, a vector network analyzer (VNA), a robot scanner, and a control computer. The open-ended rectangular waveguide (WR-62) is a near-field electromagnetic sensor and its operating frequency band is 13 GHz – 18 GHz. The VNA generates and acquires signals in the same band. The robot scanner scans the 3D structure model along the pre-designed path. The path forms two-dimensional (2D) plane according to each distance between the sensor and the target. The control computer synchronizes the devices to collect signals, and the signals are processed and stored in the computer. The system is verified through scanning the non-biomaterials such as rubber and silicon. It was shown that letters ‘I’ and ‘S’ that are made of rubber covered by silicon sheet were clearly imaged using the system. The system was also verified through scanning tissues containing porcine meat and bone. The produced 2D image were compared according to the frequency and the distance between the sensor and target. In the raw images, it was shown that the clearest image is obtained when the frequency is 14 GHz and the distance is 2 mm. To improve the quality of the image, a holographic imaging algorithm was applied to the raw data image. Then, the images were compared using the ‘contrast method’ that compares the intensity of the bone and the other parts. The result shows that the holographic imaging algorithm generates the higher quality image. It is expected that the proposed method will be more practical if the sensors are arranged to improve the scanning speed.