Loading...
Please wait, while we are loading the content...
Similar Documents
Augmenting Cartographic Resources and Assessing Roadway State for Vehicle Navigation
| Content Provider | Semantic Scholar |
|---|---|
| Author | Seo, Young-Woo |
| Copyright Year | 2012 |
| Abstract | Maps are important for both human and robot navigation. Give n a route, drivingassistance systems consult maps to guide human drivers to th eir destinations. Similarly, topological maps of a road network provide a robotic v ehicle with information about where it can drive and what driving behaviors it should use. By providing the necessary information about the driving environment, maps simplify both manual and autonomous driving. The majority of existing cartographic databases are built, using manual surveys and operator interactions, to primarily assist human navig ation. Hence, the resolution of existing maps is insufficient for use in robotics ap plications. Also, the coverage of these maps fails to extend to places where roboti cs applications require detailed geometric information. To augment the resolution and coverage of existing maps, thi s thesis investigates computer vision algorithms to automatically build lane-le vel detailed maps of highways and parking lots by analyzing publicly available carto g aphic resources, such as orthoimagery. Our map-building methods recognize image patterns and obje cts that are tightly coupled with the structure of the underlying road network by 1) identifying, without human intervention, locally consistent image cues and 2) li nking them based on the obtained local evidence and prior information about roadwa ys. We demonstrate the accuracy of our bootstrapping approach in building lane-le v l detailed roadway maps through experiments. Due to expected abnormal events on highways such as roadwork , the geometry and traffic rules of highways that appear on maps can occasion lly change. This thesis also addresses the problem of updating the resulting maps with temporary changes by analyzing perspective imagery acquired from a vi sion sensor installed on a vehicle. To robustly recognize highway work zones, our sign recogniz er focuses on handling variations of signs’ colors and shapes. Sign recognit ion errors, which are inevitable, can cause our system to misread temporary highway ch nges. To handle potential errors, our method utilizes the temporal redunda ncy of sign occurrences and their corresponding classification decisions. We demon strate the effectiveness and robustness of our approach highway workzone recognitio n thr ugh testing with video data recorded under various weather conditions. Two major results of this thesis work are 1) algorithms that a nalyze orthoimages to produce lane-level detailed maps of highways and parking lots and 2) on-vehicle computer vision algorithms that are able to recognize tempo rary changes on highways. Our maps can provide detailed information about a rout e, in advance, to either a human driver or a self-driving vehicle. While driving on hi g ways, our roadwayassessing algorithms enable the vehicle to update the resul ting maps with temporary changes to the route. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://repository.cmu.edu/cgi/viewcontent.cgi?article=1204&context=dissertations |
| Alternate Webpage(s) | https://www.ri.cmu.edu/pub_files/2012/4/ywseo-dissertation.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
