Chi N. Thai

APPLICATION OF X-RAY CT FOR DETECTION OF PHYSICAL FOREIGN MATERIALS IN FOODS

A medical X-ray CT scanner was used as a non-destructive inspection method for the detection of selected non-metallic materials embedded in various fluids and food materials. The X-ray absorption characteristics of tested materials were expressed as X-ray CT numbers. It was found that the CT number of a foreign object was affected by a systemic error (a point spread function), even if its size was greater than 1 pixel in the video image provided by the scanner. Within the interface region between the foreign object and its surrounding matrix, CT numbers were found to vary smoothly between characteristic values for the object and the matrix thus pointing to important effects to consider when developing X-ray CT techniques for the detection and identification of foreign materials in foods. Specific threshold values for CT numbers of selected foodstuffs were defined for successful detection of foreign materials. This study also pointed out the limitation of using X-ray linear absorption coefficients from typical CT systems to detect certain combination of foreign materials and foodstuffs.

A Genetic Algorithm Based Tool for Analysis and Modeling of Multi-Spectral Data

This report describes the development of statistical models to determine the relationships between vegetation indices and bush bean plant nitrogen stress levels using 2 approaches, statistical analysis with exhaustive search by SAS procedures and optimal search using genetic algorithm (GA). With SAS, polynomial regression was used to fit data under all possible twowaveband combinations. The best wavebands identified were 700nm, 710 nm, 720 nm, and 750 nm. The best model selected by both SAS and GA was a 2nd degree polynomial for RVI (= R e2 / R e1) with an adjusted R2 value of 0.9144. Comparison of these two approaches showed that GA integrated both statistical analysis and model selection and was as accurate but more efficient than SAS.

Robotics-based freshman immigration course into computer systems engineering

A robotics-based freshman immigration course into Computer Systems Engineering (CSE) was created to introduce students to the main design issues in CSE: multi-controller programming, sensor interfacing, servo actuation and serial communications. Freshman students were found to be enthusiastic about the Robotis Bioloid system used and found it challenging but not overwhelming. The instructional materials were designed around 3 projects: car bot in maze, GERWALK bot penalty kick, and GERWALK bot negotiating stair steps. Mid-term course assessment results are presented in this article.

Integration of Tablet PCs and Multimedia Technologies in Engineering Instruction

This is a report about a continuing project for developing instructional technologies and materials for in- and outside- of-classroom activities to promote student learning. In-class activities included student-teacher interactions mediated via NetSupport Manager, student hypermedia note-taking via Windows Vista Snipping Tool and Camtasia, algorithms development and quizzes correction using Digital Pen Technologies. Outside-of-class activities included reviewing narrated Flash movies for homework solution design hints and for pre-recorded narrated guidance during phases of a semester project using robots equipped with Machine Vision capability. This facility had been used to teach Java Programming and Systems Simulation for 3 semesters. Student feedback and instructor assessments regarding the effectiveness of technologies used are reported herein.

Developing Active Learning Activities for Computer Programming and Machine Vision Courses

This is a report about a continuing project for developing instructional technologies and materials for in- and outside- of-classroom activities to promote student learning. In-class activities included student-teacher interactions mediated via NetSupport Manager, student hypermedia notetaking via Silicon Chalk, algorithms development and quizzes correction using a Tablet PC. Outsideof- class activities included reviewing narrated Flash movies for homework solution design hints and for pre-recorded narrated guidance during phases of a semester project using robots equipped with Machine Vision capability. This facility had been used to teach Java Programming, Systems Simulation and Applied Machine Vision courses with overall favorable assessments from students. Student feedback and instructor assessments regarding the effectiveness of technologies used are reported herein. Particular student assessments regarding Tablet PC effectiveness and value of the robot projects are also presented.

Software Design for Dynamic Stitching of Multi-Spectral Images of Field Crops

We proposed an alternate configuration for dynamic multi-spectral imaging of plants at ground level, using a one-sensor monochrome FireWire-A camera combined with a liquid crystal tunable filter (LCTF) tunable at 760 nm and 695 nm. We developed an algorithm based on the traditional NDVI procedure to untraditionally solve the problem of image correspondence within sequences of spectral images collected from a moving platform. This algorithm was designed from a multi-threaded software engineering approach and had been shown to work well during acceleration, steady, and deceleration phases of a moving platform under laboratory conditions and one limited field test. The algorithm had two components: one to determine the optimal number of pixel rows passing by the camera viewport during the time period needed for the LCTF to switch between wavelengths; the other to determine the proper conditions to trigger an image saving event so as to minimize the number of images recorded that would however have enough overlaps to permit their assembly into potential field-level maps. The proposed system could process up to 14 image sets per second (3 images per set−760 nm, 695 nm and NDVI) and was found to tolerate light breezes under field conditions.

Teaching Robotics to Students with Mixed Interests

An innovative and flexible course in robotics was created to accommodate the interests of engineering and non-engineering students at UGA. It sought the integration of multi-threaded programming, advanced sensors interfacing (video camera and balance sensor), actuators control and wireless communications (Zigbee). It was based on the Bioloid Expert and Premium robotics systems from Robotis, Inc. Details of course design and materials, and of student projects are described herein.

A Dual-Mode Collaborative Teaching and Learning Classroom Environment

As “hand-written notes” and “personal narrations” are the most natural tools to record, review, integrate and develop our own learning process, this article describes a collaborative learning and teaching environment integrating the use of Pen Tablets and multimedia technologies at the student/teacher level with the goal of achieving active learning for students in the classroom. An innovative hardware and software integration of the PCI NetSupport Manager Suite, TechSmith Camtasia Studio and Wacom Pen Tablets was implemented to provide both “teacher-centric” and “student-student collaborative” modes in this classroom. As this project was just implemented in mid-August 2007, only preliminary assessment results are presented in this article.

Development of an LCTF/FireWire Spectral Imaging System

We proposed an alternate configuration for dynamic multi-spectral imaging using a onesensor monochrome FireWire-A camera combined with a discretely tunable LCTF and developed an algorithm based on the traditional NDVI procedure to untraditionally solve the problem of image correspondence within sequences of spectral images collected from a moving platform. This algorithm had been shown to work well during acceleration, steady and deceleration phases of a moving platform under laboratory conditions.

Multi-Threaded Design of Spectral Imaging Software

In order to elicit the maximum performance of a multi-spectral imaging system designed for plant health mapping under field operating conditions, a multi-threaded approach was used in creating software modules to interface and integrate the use of a FireWire_A camera capable of 100 fps at 640x480 pixels (8-bit gray levels) and a Liquid Crystal Tunable Filter having a cycle time of 17 ms to switch between 692 and 755 nm. Due to the discrete-event nature of the multi-spectral imaging system, a multi-threaded software package was developed to distribute the workload of camera control, calibration, image capture, and image analysis over multiple worker threads implemented in the Microsoft Foundation Classes. The implementation of a priority-based, multi-threaded system achieved maximum efficiency by minimizing system downtime during processes that are timeintensive but computationally simple. In this system, computationally expensive routines such as image analysis are preformed while image capturing routines are suspended awaiting an event that is independent of the system.