Design Plan
Design Concept Narrative Summary
Our main design goal for this project is to develop a dual-camera demonstration kit for Micron which utilizes stereo input to detect the approach of an obstacle and in response, outputs an audio or visual alarm. This will be implemented by writing algorithms to obtain position and velocity information (i.e., 3D information) by analyzing the differences between parallax images taken by our dual-camera setup.
Design Diagrams
Below are diagrams of how our proposed system will work. Our methods for obtaining position and velocity information are currently under research.
Proposed Camera Layout Diagram for 3-D Detection
System Function Diagram
- An object approaches the stereo imaging device.
- The dual cameras set a linear distance apart so as to import parallax images pick up movement and send output to the processor/PC.
- The processor evaluates the data sent by the cameras to detect movement and determines if the object is close enough or moving fast enough to cause alarm.
- The processor will send an output signal which activates the audio/visual device if an alarm is deemed necessary.
Design Philosophy
To meet the needs of the client, our design philosophy is to remain aware of the impact of our design choices in many different areas. Little has changed in this philosophy since the last report.
- Costs – Micron is supplying all of the necessary equipment.
- Code – Needs to be well documented and commented for each phase so that it can be adapted to fit new requirements, both later in this project and for other Micron projects.
- Environment – Needs shielding from interference and static electricity.
- Hardware Maintenance – None required.
- Reliability – Needs to respond accurately without excessive false positives.
- Manufacturing – Hardware is already created and supplied by Micron.
- Packaging – Needs to be transportable.
- Performance – Needs to work in real time. Needs to be accurate, and not respond to false positives.
Goals and Subsystems
Our goals define five processes that must function together to achieve our desired result. Processes 3 – 5 have been updated since the last status report to reflect changes in the requirements and objectives.
- Bring in a pair of simultaneous images.
- Process the images.
- Calculate the placement and closing speed of objects.
- Determine whether an object is a danger; if yes, activate alarm.
- Analyze implementation on an FPGA (if there is time remaining).
Team Organization
Our individual responsibilities have been divided according to the major areas within the project. Each person’s responsibilities are tentative depending on the constraints of time, the project, and/or team. Each responsibility represents an individual’s emphasis and in no way limits the scope of his or her work.
- Software – (Jose Madrid)
- Hardware – (James Vun Kannon)
- Algorithms – (Lauren Olson)
- Research – (Kimberly Adler)
Analysis and Simulations
We are currently in the process of analyzing simulated images to develop algorithms for identifying objects and determining the z-distance of the objects. These images are being simulated using a single camera to take two pictures of an object.
Tradeoffs and Challenges
We anticipate dealing with the following challenges:
- Creating an algorithm to identify the distance of a generic object.
- Focusing the cameras.
- Ignoring non-hazardous, background objects.
- Selecting accurate object identification algorithms that can yield quick, clear results without overburdening the hardware.
- Real time imaging requires fast processing and may limit FPGA implementation.
We will make decisions by researching/developing possible solutions and creating design matrices to choose between them.
Parts and Research
We are using the parts specified by the client for demonstration. The only part remaining to be chosen is the mounting for the pair of cameras. These will require further research into the exact separation distance of the cameras and the vibration tolerances of the algorithms.
Constraints
- Cost - Cost is not a significant issue for our team, as all relevant purchases and hardware will be made and supplied by Micron. Mounting hardware will not contribute significantly to the cost of a demo kit.
- Environment - Our project will not create interference harmful to the environment or wireless communications.
- Manufacturability – The final demo kit will be easy to assemble. Camera hardware is manufactured and supplied by Micron.
- Health and Safety – This project, if modified for use as a safety device on a vehicle, will enhance the safety of the driver and sentient obstacles behind the vehicle by warning the driver if a collision is eminent. Our demo kit must not pose a significant health or safety threat to the user.