Fred W. DePiero

Surface Registration at 10Hz Based on Landmark-Graphs: Benefits for a Scalable Remote Viewing System

Fred W. DePiero — Tue, 14 Oct 2008 12:32:32 PDT

Real-time surface registration is a key technology for the development of future remote viewing systems. An architecture for a video distribution system supporting multiple users, with individual viewpoint selection, is suggested. The approach would provide a transmission bandwidth independent of the number of users, for scalability. The proposed architecture uses a method of surface registration based on landmark-graphs. Results from 141 test trials on synthetic scenes indicate that a mean absolute positioning accuracy under 1% of the sensor field of view is possible. The mean rate for registration was 10Hz, with a standard deviation under 10%. Tests were benchmarked on a 900MHz PC. The sensor images were 200x200 pixels and contained both range and color imagery.

Improving ABET Assessment Processes at Cal Poly

Fred W. DePiero — Thu, 14 Aug 2008 11:59:35 PDT

Real-Time Range Image Segmentation Using Adaptive Kernels and Kalman Filtering

Fred W. DePiero — Wed, 06 Aug 2008 13:35:18 PDT

Segmentation is a fundamental process affecting the overall quality and utility of a machine vision system. Range Profile Tracking (RPT) is a systematic approach for stable, accurate and high speed segmentation of range images that is based on Kalman filtering. Tests of RPT have produced stable decompositions of second order surfaces bounded by jump and crease discontinuities, having a volumetric error of a few percent, in under 6 sec. for a wide variety of conditions. Results from over 900 tests on synthetic scenes and 150 real range images are presented.

Reduced Bandwidth for Remote Vehicle Operations

Timothy E. Noell — Wed, 06 Aug 2008 13:35:00 PDT

Oak Ridge National Laboratory staff have developed a video compression system for low-bandwidth remote operations. The objective is to provide real-time video at data rates comparable to available tactical radio links, typically 16 to 64 thousand bits per second (kbps), while maintaining sufficient quality to achieve mission objectives. The system supports both continuous lossy transmission of black and white (gray scale) video for remote driving and progressive lossless transmission of black and white images for remote automatic target acquisition. The average data rate of the resulting bit stream is 64 kbps. This system has been demonstrated to provide video of sufficient quality to allow remote driving of a High-Mobility Multipurpose Wheeled Vehicle at speeds up to 15 mph (24.1 kph) on a moguled dirt track. The nominal driving configuration provides a frame rate of 4 Hz, a compression per frame of 125:1, and a resulting latency of ~1s. This paper reviews the system approach and implementation, and further describes some of our experiences when using the system to support remote driving.

Waste-Surface Mapping of the Fernald K-65 Silos Using a Structured Light Measurement System

B. L. Burks — Wed, 06 Aug 2008 13:34:38 PDT

During FY 1991, a remotely operated surface-mapping measurement system was developed by the Robotics & Process Systems Division at Oak Ridge National Laboratory for use in the K-65 waste-storage silos at Fernald, Ohio. The mapping system used three infrared line-generating laser diodes as illumination sources and three high-resolution, low-lux, calibrated, black-and-white, chargecoupled- device video cameras as receivers. These components were combined to form structured light source range and direction sensors with six different possible emitter-receiver pairs.A technology demonstration and predeployment tests were performed at Fernald during July and August 1991, using the empty Silo 4 into which was placed rectangular objects of known dimensions. These objects were scanned by the structured light sources to demonstrate functionality and verify that the system was giving sufficiently accurate range data in three dimensions. The tests provided a checkout of installation and operational logistics, prior to field deployment, as well as verification of accuracy and repeatability.In September and October 1991, the structured light sources were then deployed in Silos 1 and 2 to scan the waste surfaces. The resulting data were merged to create three-dimensional maps of those surfaces. A bentonite clay cap was placed over the waste surfaces in November 1991. Surface maps were obtained in December 1991 of the bentonite clay caps. The change in surface height before and after bentonite addition was utilized as a measure of clay cap thickness. Verification of clay cap thickness was required to support a Department of Energy (DOE) Removal Action Milestone defined in the Amended Consent Agreement with the Environmental Protection Agency under Comprehensive Environmental Response, Compensation, and Liability Act Section 106 and 120.This report includes final results of the surface mapping campaign as well as a description of system components. Mapping system sensors and control hardware are described. Calibration and data acquisition methodology are described as well as the operator interface that facilitated these operations. The sensor system development was sponsored by the DOE Office of Technology Development, Robotics Technology Development Program. Field deployment in the K-65 silos was funded by the DOE Office of Environmental Restoration.

SIPTool: The 'Signal and Image Processing Tool' An Engaging Learning Environment

Fred W. DePiero — Wed, 06 Aug 2008 12:00:06 PDT

The 'Signal and Image Processing Tool' is a multimedia software environment for demonstrating and developing Signal & Image Processing techniques. It has been used at Cal Poly for three years. A key feature is extensibility via C/C++ programming. The tool has a minimal learning curve, making it amenable for weekly student projects. The software distribution includes multimedia demonstrations ready for classroom or laboratory use. SIPTool programming assignments strengthen the skills needed for life-long learning by requiring students to translate mathematical expressions into a standard programming language, to create an integrated processing system (as opposed to simply using canned processing routines).

NetExam: A Web-Based Assessment Tool for ABET2000

Fred W. DePiero — Wed, 06 Aug 2008 12:00:01 PDT

Abstract ⎯ NetExam is a web-based testing engine. In addition to automated testing and grading capabilities, NetExam computes statistics that are tied directly to program outcomes, for ABET2000 assessment purposes. NetExam provides advantages over scantron-style testing as it also presents statistics on program outcomes on the web. This facilitates review and black board-style comments by program constituents. Also, all of the exam generation, grading, statistics and program assessment features are integrated into the web-based system. First usage is planned for the 2001-2002 academic year.

Structural Graph Matching With Polynomial Bounds On Memory and On Worst-Case Effort

Fred W. DePiero — Wed, 06 Aug 2008 11:59:58 PDT

A new method of structural graph matching is introduced and compared against an existing method and against the maximum common subgraph. The method is approximate with polynomial bounds on both memory and on the worst-case compute effort. Methods work on arbitrary types of graphs and tests with strongly regular graphs are included. No node or edge colors are needed in the methods; the common subgraph is extracted based in structural comparisons only. Monte Carlo trials are benchmarked with 100% additional (clutter) nodes. Results are shown to be typically within 1-2 nodes of the maximum common subgraph. Over 7500 test trials are reported with graphs up to 100 nodes.

Remote Driving With Reduced Bandwidth Communication

Fred W. DePiero — Wed, 06 Aug 2008 11:59:54 PDT

Oak Ridge National Laboratory has developed a real-time video transmission system for low-bandwidth remote operations. The system supports both continuous transmission of video for remote driving and progressive transmission of still images. Inherent in the system design is a spatiotemporal limitation to the effects of channel errors. The average data rate of the system is 64,000 bits/s, a compression of approximately 1000:1 for the black and white National Television Standard Code video. The image quality of the transmissions is maintained at a level that supports teleoperation of a high-mobility multipurpose wheeled vehicle at speeds up to 15 mph on a moguled dirt track. Video compression is achieved by using Laplacian image pyramids and a combination of classical techniques. Certain subbands of the image pyramid are transmitted by using interframe differencing with a periodic refresh to aid in bandwidth reduction. Images are also foveated to concentrate image detail in a steerable region. The system supports dynamic video quality adjustments between frame rate, image detail, and foveation rate. A typical configuration for the system used during driving has a frame rate of ~ 4 Hz, a compression per frame of ~ 125:1, and a resulting latency of < 1s.

Multidisciplinary Experiences for Undergraduate Engineering Students

Fred W. DePiero — Wed, 06 Aug 2008 09:59:43 PDT

Multidisciplinary (MD) team skills are of increasing importance in industry, and are required for programs that are ABET accredited. This can be challenging to provide, particularly for programs with high unit counts and a large student body. We have responded by establishing a multidisciplinary graduation requirement across the College of Engineering at Cal Poly. This requirement is mandatory. It may be satisfied by various curricular and co-curricular routes, easing throughput issues compared to having a single venue. In the paper we describe activities that satisfy the MD requirement, our general approach and assessment methods.