As part of ITS technology, to achieve quick map updates, we propose a method for automatically detecting changes in streetscapes from images captured by car-mounted omnidirectional cameras. It comprises two stages; accurate alignment of a map and street images taken at various times, and detection of changes in streetscapes from the aligned data. The system will collect data via many free-running cars fitted with low-cost equipment to obtain images at various times and along routes. In the first stage, we process the alignment of the image frames taken at same locations and determine the accurate position information of each frame by a method composed of dimension reduction and DP matching. Then in the second stage, we detect changes in streetscapes from images taken at various times. Experiments with 44 data items which were collected over about a year, demonstrate the effectiveness of our method
[1]
Edward Verbree,et al.
INTERACTIVE NAVIGATION THROUGH DISTANCE ADDED VALUED PANORAMIC IMAGES
,
2004
.
[2]
Yasushi Yagi.
Real-time Omnidirectional Image Sensors
,
1995
.
[3]
Xiaolong Wang,et al.
Pervasive views: area exploration and guidance using extended image media
,
2005,
ACM Multimedia.
[4]
Hiroshi Murase,et al.
Visual learning and recognition of 3-d objects from appearance
,
2005,
International Journal of Computer Vision.
[5]
H. Kawasaki,et al.
A photo-realistic driving simulation system for mixed-reality traffic experiment space
,
2005,
IEEE Proceedings. Intelligent Vehicles Symposium, 2005..