| 摘要: |
| 钎焊环检测系统中，采集的图像序列可能具有不完整环或重叠环。通过调整传送带速度和设置固定相机的曝光时间，确定了2幅图像的重叠区域；将图像去噪、灰度化和二值化，得到了具有清晰轮廓的图像；然后结合LSSVM和CA对图像边缘进行了检测，利用互信息法实现了图像配准；最后采用渐入渐出法进行了图像融合，得到了最终的拼接图像。实验表明，这种方法可以快速准确地实现钎焊环图像的拼接。 |
| 关键词: 钎焊环 最小二乘支持向量机 元胞自动机 图像配准 图像融合 |
| DOI： |
| 分类号:TS865 |
| 基金项目: |
|
| Study on Image Mosaicing Methods of Brazing Rings Inspection |
|
ZHONG Xing, XU Xiang-hong
|
|
China Jiliang University, Hangzhou 310018, China
|
| Abstract: |
| In brazing rings inspection system, there may be incomplete or overlapping rings in image acquisition sequence. The overlapping area of two images was determined by adjusting the speed of conveyor and exposure time of fixed camera. Then, LSSVM and CA was applied to detect the image edge and image registration was implemented by cross correlation method. The final image was acquired by image fusion based on fade-in-fade-out method. The experiment results indicated that this method can achieve brazing rings image mosaicing quickly and accurately. |
| Key words: brazing rings least squares support vector machine (LSSVM) cellular automata (CA) image registration image fusion |
