3种不同模型对木材表面缺陷图像分割算法的比较

Contrast of Three Image Segmentation Algorithms for Wood Surface Defects

木材表面缺陷会严重影响木材的质量和使用价值,因此对木材表面缺陷图像分割的研究有利于提高木材的利用率。本文分别对红皮云杉含有虫眼、活节、死节3种典型木材缺陷的图像采用改进的C-V模型、改进的GVF Snake模型和改进的GAC模型进行分割试验,对3种改进算法的复杂程度、分割时间、分割结果的完整性以及抗噪性进行对比和分析。结果表明,改进的GAC模型算法较为优越,其分割算法简单,运行时间短,缺陷分割效果较好,抗噪性强。而改进的C-V模型算法、改进的GVF Snake模型算法的分割效果和抗噪性最差,不宜作为3种木材表面缺陷图像的分割算法。

Since the surface defect of the wood can seriously affect the quality and the use value of the wood, the research of image segmentation algorithm for wood surface defects can increase the utilization rate of wood. This pa-per respectively research three typical wood defects with wormhole, living scab and dead scab. For segmentation test, the improved C-V model, the improved GVF Snake model and the improved GAC model were used, and the complexity of the three improved algorithms, the time required for segmentation, the integrity of segmentation re-sults, and the noise performance of the algorithm were analyzed through the test. The results showed that the im-proved GAC model was superior to other methods since its segmentation algorithm was simple, its running time was shorter, its defect segmentation results were complete and its noise immunity performance was significant. The im-age segmentation result and noise immunity of the improved C-V model and the improved GVF Snake model were not significant.