Loading...
Please wait, while we are loading the content...
Similar Documents
Infrared Night Vision Based Pedestrian Detection System
| Content Provider | Semantic Scholar |
|---|---|
| Author | Fang, Chiung-Yao |
| Copyright Year | 2008 |
| Abstract | When people drive in low-visibility condition at night, even though the cars are equipped with head lights, they’re often unable to provide enough lighting to locate suddenly-appearing pedestrians. Hence, we have investigated in this system which utilizes the combination of computer-vision techniques and computational analyses, to detect the location and moving direction in advance, which could then warn any potential danger to reduce the chances of accidents. Unfortunately, the analysis of normal video-taping at night has its limits, such as the noises produced by electronic devices, even if the aperture and ISO are large enough. However, even though visible light is dim during night time, infrared is emitted whenever an object with considerable temperature is present. In this case, human body temperature is relatively higher than the environment and thus the more intense infrared can be easily detected. After applying the computation of spatial distance, shape, the intensity of infrared and appropriate position, we might infer the objects which satisfy these analysis to be potential pedestrians. Moreover, because of the fast-speed driving (especially on highways), this system has to be efficient and gives real-time feedback to be considered useful. PREVIOUS STUDIES AND DOCUMENTATION The idea of using infrared images to detect pedestrians was introduced by M. Nilson et. al, |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.csie.ntnu.edu.tw/~u92210/computer_vision_paper.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
