Objectives
1. Review of Predictability Attributes.
2. Image Processing Approach.
3. PACC Analysis Clarification.
| Date of Meeting | Start Time | End Time |
| 04/03/2009 | 14:00 | 15:00 |
| Meeting Called By | Meeting Location | |
| VdashNeg | SEI 3505 | |
| Facilitator | Scribe | Timekeeper |
| Saurabh | Shigeru | Jonathan |
| Attendees | ||
| Scott Hissam | ||
| Dan Plakosh | ||
| Marsha Pomeroy-Huff | ||
| Grace Lewis | ||
| Berin Babcock-McConnell | ||
| Saurabh Gupta | ||
| Jonathan Hartje | ||
| Shigeru Sasao | ||
| Sidharth Surana |
| Topics | |||
| Time | Topic | Discussion Leader | Desired Outcome |
| 14:00 | 1. Review Predictability Attributes | Shigeru | A common understanding of the predictability scenarios. |
| 14:30 | 2. Image processing approach | Saurabh, Berin, Jonathan | Discussion about the image processing approach of the team. |
| 14:30 | 3. Clarification about PACC | Sidharth, Shigeru | Clarification of some of the problems the team has been facing wrt performance and behavior analysis. |
We would like to propose that we discuss the image processing approach with Dan while Scott can clarify our doubts about PACC behavior and performance analysis simultaneously.
1. Review Predictability Attributes
1.
2. Image processing approach
1. We have been assuming a level field of operation for the robot. This means, we are not expecting the robot to travel over any object (such as a cord protector for example) or travel up and down ramps. This assumption has an impact on the assumptions we can make in selecting and creating our image recognition algorithms.
2. Dan mentioned Open CV. Open CV is an image processing kit? library? initiated by Intel but now available as open source.
3. Dan suggested we look into books titled something like "Graphics Gems". These books have code snippets that may be useful for us.
4. Dan noted that image processing can be a bit of an art in selecting threshold, etc, values. The important part is to collect data and see what works.
3. Clarification about PACC
1. keyboard eventDistribution issue clarified. (eventDistribution as 0 written in Support/machdep/CoreServices.ccl)
2. stack overflow issue clarified.
4. Action Items
1. Shigeru to create simple code to demonstrate issue with annotating instance vs component sink pins.
2. Shigeru to create simple code to demonstrate watch dog error.
3. Sid/Shigeru to Change component to pass in IP, port.
