A little about IPRO 324

Welcome to the online home of IPRO 324. Spring '03 is the third run of this Inter-professional Project (IPRO) at the Illinois Institute of Technology. IPROs are semester-long undergraduate projects based on real world topics from sponsors that involves students from multiple disciplines across the university.

Located in Elmhurst, Illinois, The Chamberlain Group, Inc. is the world's largest manufacturer of residential garage door openers and commercial and gate operators for use on cars, boats and planes. Chamberlain leads the industry in product innovation and quality, and is the low-cost producer. At this time, through its Advanced Development group, Chamberlain is interested in exploring new devices that can enhance the functionality of its garage door operator systems. Residential garage door operators currently utilize an IR emitter and detector pair to identify the presence of an object breaking the "beam". This sensor is used to reduce the risk of a moving garage door striking an object in its path. The IPRO team will investigate the use of an inexpensive CCD camera to replace, and potentially expand, this functionality.

This being the third and final semester, we would like to give you a brief overview of the research that was accomplished and the progression from theory to practicality term by term. Keep in mind that the primary focus of this IPRO was on the feasibility of the product for commercial use.

Semester 1: Spring 2002

Image Processing: Pattern Recognition

This team created working software to process an image taken from a web cam to determine if an object was in the way of the garage door. In their pattern recognition algorithm they assumed that the pattern was always in a fixed position in the image and that the environment was ideal (eg. ideal lighting conditions). A cropped image of the pattern was used to detect an obstruction in a given frame. The image processing that occurred looked for the edge of the cropped image in the frame regardless of the pattern itself. Although this method allowed for a practical real-time assessment of the garage door functionality (an image was processed frequently) it was obviously not practical. Correcting issues such as translation of the camera, scaling due to varied garage door widths, lighting conditions, and required camera resolution was not well explored.

Summary: There was not much documentation or evidence to determine the feasibility of such a product commercially but this team laid the groundwork for the next two semesters. They did not use the CCD camera but developed a functional product with a web cam and were able to show how the hardware and software could be integrated to develop a working model. Many of the ideas that they had were used again or provided a minimum to improve upon for later groups. This team discovered some of the major obstacles that must be overcome in order to develop a commercial product.

Semester 2: Fall 2002

Image Processing: Pattern Recognition

This group opted to continue the use of a web cam for simplicity (the software for frame grabbing was ready to use from the start). Their chief goal was to correct the issues of lighting, scaling, and skewing of the image which were the major drawbacks in the previous semester. As in the previous IPROs a cropped rectangular image of the pattern was used to identify obstructions. The pattern recognition software was able to correct for very small skewing and scaling problems. Also, an effort was made to correct for lighting conditions which was marginally successful. The software was much slower than the previous group although various issues were addressed.

Improvements for the Next Term: This group created a simple flow-chart of the image processing software which served as a template for our group. Also, they prepared a clearly documented analysis of recommended improvements for the future IPRO. These improvements were: 1. The use of a CCD camera for both enhanced resolution and speed 2. Advisement of the continued use of Matlab for image processing functions 3. Advisement to convert Matlab code to C++ for the purpose of speed 4. The proposition of a circular pattern to eliminate skewing This group also created a physical model of the garage door for demonstration purposes.

Summary: The analysis of improvements required for this product to be successful was an important addition made by this IPRO group. All of the recommendations listed above were used in the following semester. As before, there was not enough data to determine the feasibility of the product for commercial use.

Semester 3: Spring 2003

Image Processing: Pattern Recognition

The pattern recognition software is written entirely in Matlab, while the C++ code was written to capture the image from the CCD camera, run the Matlab program, and provide a user interface. This semester the skewing problem was eliminated and the scaling issue was improved enough to be practical. The algorithm itself, however, is too slow for use in this application. Much of the lighting conditions are solved with the addition of the CCD camera but extreme lighting conditions are still a problem. Other issues such as the required resolution for image processing have been determined and documented in the feasibility report.

Feasibility: We were able to somewhat accurately determine the feasibility of this product for commercial use. More documentation may be found within the final report.

Summary: We have come to the conclusion that this product is not yet ready for commercial use. This is mainly due to the cost associated with assembling a safe and efficient product using CCD camera technology. The determination of this fact was a long and drawn out process with each IPRO group contributing their fare share. Please browse through this website to view official reports and documentation.