牛胸椎骨图像中软骨和硬骨区域的自动分割技术

Auto-segmentation of cartilage and bone areas in thoracic vertebrae images

研究了基于Ohta颜色空间的牛胸椎骨图像中软骨和硬骨区域自动分割技术,这为应用计算机视觉评定牛肉的生理成熟度提供了条件。首先使用数码照相机现场拍摄牛胸椎骨RGB格式图像传输到计算机内,并转换到Ohta颜色空间,然后对Ohta颜色系统的第二分量(I2)和第三分量(I3)图像进行二值化、形态腐蚀操作、扫描填充、面积阈值分割等一系列处理。这样,可获得牛胸椎骨图像中软骨和硬骨的区域。研究结果表明,图像处理方法自动分割牛胸椎骨图像中的软骨和硬骨是可行的,为牛生理成熟度的计算机视觉自动检测打下了基础。

The automated segmentation of the cartilage and bone areas in the thoracic vertebrae images based on OHTA color systems was discussed in this paper. The thoracic vertebrae images were captured with digital camera and were saved in computer. The image in RGB format was first transformed into OHTA color system; then I2 and I3 values of OHTA system were effective in segmenting cartilage and bone areas, respectively. After applying binary image, morphological erodes, hole filling and area threshold on I2 and I3 images, the geometric properties of cartilage and bone were obtained. The results of this work contribute to the automatic computer vision system for beef maturity detecting and grading.