基于照度-反射模型的在线水果表面缺陷检测

水果表面缺陷是影响水果价格最有力的因素之一，目前大部分研究基于静态图像采用较复杂的算法分割水果表面缺陷，因此很难满足在线水果缺陷的快速检测。本文介绍了一种在线水果表面亮度均一化校正及单阈值缺陷分割方法。首先，通过RGB颜色空间转换获取HIS空间图像的色调（H）分量，基于H分量建立掩模对RGB图像的R分量执行掩模去背景；然后，基于照度-反射模型，利用低通滤波获取R分量图像的亮度分量，利用此亮度分量对去背景后的R分量图像进行亮度均一化校正；最后，利用一个简单的阈值对亮度校正后的图像进行缺陷分割。利用此算法，对416幅图像的检测结果表明总体检测率超过99%。该方法简单、有效，在在线水果缺陷检测中具有较大的应用潜力。

Defects detection of fruit in real-time based on illumination-reflectance model

The presence of skin defects is one of the most influential factors in the price of fruit. Most of the current researches used static images and more complex algorithm to segment defects on fruit surface. Therefore, it was very difficult to meet defects detection of fruit in real-time. An effective algorithm was proposed to homogenize illumination on fruit in order to segment defects using a single threshold value. Firstly, hue component was obtained by changing RGB colour space to HIS colour space. The mask was created based on hue component and used to remove R-component image background. Then, based on illumination-reflectance model, the illumination component was extracted from R-component image by low pass filtering. The illumination component was used to homogenize illumination on R-component image. Finally, defects were successfully segmented at one time by subjecting the corrected image to a single threshold value. The experimental results with an over 99% recognition rate based on 416 images showed that the proposed algorithm was simple and effective and promoting to apply in real-time detection of defects on fruit.