Applications of region developing feature in wheel image segmentation

作者：Yang Jingyu^*; Wang Song; Du Xiaogang

来源：Journal of Theoretical and Applied Information Technology, 2012, 46(2): 942-947.

• Area ratios
• Base points
• Contour image
• Dynamic detection
• Image characteristics
• Image features
• Image segmentation algorithm
• Image verification
• In-wheel
• Measurement techniques
• Misclassification error
• Noise immunity
• On-line detection
• Screenshots
• Seed point
• Segmentation results
• Single connection
• Traffic safety
• Tread wear

Online detection of protection orbit round as an important part of the train, the tread wear parameters traffic safety is important. Like the wheel of the measurement techniques based on optical screenshots Dimensions dynamic detection ′access to a complete round of the contour image is essential. For the round image features is proposed based on region developing the round image segmentation algorithm. Round of the algorithm according to the different circumstances of the image characteristics through the base point, select the appropriate seed points and determine the appropriate growth criteria. After region developing, effectively extract the wheel contour image. Through a large number of the image verification, the segmentation results of the algorithm figure overlap area ratio is greater than 82%, the misclassification error area ratio of less than 0.01%. Round extract can be effective in a variety of circumstances of the contour curve has good noise immunity.

• 出版日期2012