Bridging the gap between Computations and Experiment in Electromagnetic Inversion

点击翻译按钮获取中文摘要

The ability to see things that are normally hidden from the naked eye is an enormous benefit to humankind. Consider medical imaging (detecting and monitoring disease) or geo-prospecting/monitoring (detecting economically valuable natural resources). All these benefits require being able to `see' inside of normally inaccessible regions, like the human body or underground. Electromagnetic inversion uses electromagnetic (e.g. radio) waves and imaging algorithms to see inside a wide variety of objects. Images are created using data-collection hardware (antennas and radio transceivers), computer models, and optimization algorithms. Importantly, these images are quantitative with each pixel having a number associated with the physics of the material. Applications of electromagnetic inversion are extremely broad: it can be used to image targets such as breast cancer, stored grain, as well as oil-fields. Imaging these diverse targets occurs with virtually identical processes and algorithms, thus this research program proposes fundamental advances that will apply to multiple socio-economic and environmental needs. Electromagnetic inversion uses more mathematically complex algorithms to image interiors of objects than some other imaging technologies (e.g. x-rays), but for many applications (e.g. grain bins) it remains the only way to generate information about target interiors. Electromagnetic inversion presents a challenging research problem, but its potential has made it an active area of research. The long-term vision of my proposed research program is to create electromagnetic inversion systems and associated software to image the inside objects and regions where such information is valuable. The research program will take an engineering approach to imaging (a) stored-grain in order to maximize crop value and minimize losses due to spoilage, (b) biomedical imaging, particularly the detection and monitoring of breast cancer. The barriers facing the deployment of such imaging systems are solvable through good engineering research, and these technologies will be made practical by: (1) reducing modeling error - the difference between the actual system and the computational model of the system; and (2) advancing the inversion algorithm workflows that generate the images from these systems. This research program will advance imaging systems monitoring Canadian grain and create a series of prototype biomedical imaging systems with higher resolution than existing technologies. Hardware, software, and algorithms will all be developed, but the main novel contributions will be in the systems-level integration, workflows and calibration. The benefits of solving these problems are clear: higher quality (and more) grain for Canadian producers and exporters, or the ability to image cancers with non-ionizing radiation. The benefits to Canada will include research advances, HQP training, direct employment, larger exports, and development of new technologies.