Design of an Ultra-wideband Frequency System for Non-destructive Root Imaging

Abstract

This thesis outlines the design and implementation of an ultra-wideband imaging system for use in imaging potted plant root system architectures. Understanding the root system architecture as plants develop is critical for plant phenotyping and ultra-wideband imaging systems have shown potential as a portable, low-cost solution to non-destructively imaging root system architectures. The proposed system is separated into three main subsystems: a Data Acquisition module, a Data Processing module, and an Image Processing and Analysis module. For each module, essential parameters and variables which largely affect the quality of the produced images and measurements of the system are analyzed and discussed.

The Data Acquisition module is responsible for collecting ultra-wideband signal reflections off the potted roots in dry soil. The most critical variables for performance of the entire system are the relative permittivities of the root and the soil. Insufficient contrast between root and soil relative permittivity results in poor performance of the imaging system. Both simulated (using finite-difference time-domain methods) and experimental trials were performed and designed for data collection. The Data Processing module receives the ultra-wideband reflection data from the Data Acquisition module and produces a 2D image using delay-and-sum beamforming. This method takes advantage of known physical and electrical parameters to generate an energy mapping of reflective objects in the soil medium to be imaged. Careful design of parameters such as the steering vector and window size are essential to optimizing the quality of the results.

The Image Processing and Analysis module removes any artifacts present in the produced images from the Data Processing module by primarily using morphological transformations. A modified top-hat transformation is used and the size of the structuring elements help remove unwanted artifacts.

The system performs reasonably well under controlled soil conditions, and there are large improvements to be made with increasing the bandwidth of the ultra-wideband device. However, since the performance of the device is extremely reliant on the soil conditions, it is recommended that further work on ultra-wideband imaging systems for roots to be focused on measuring and modeling the complex electromagnetic properties of soil at high frequencies.