基于机器视觉的荔枝果实采摘时品质检测技术

为了在荔枝采摘时实时判断果实的品质状态,通过分析自然环境中荔枝不同生长期的图像,对荔枝果实未成熟、成熟、成熟后外表腐烂变质的3种情况进行了图像数据分析。选取了YCbCr颜色模型,利用探索性分析法对荔枝不同部位、不同光照、不同生长期的荔枝图像的Cr分量进行了数据分析与统计,确定了辨识荔枝果实未成熟与成熟的Cr分量的阈值范围;对于成熟的荔枝,采用边缘提取与Hough圆拟合方法对其Cr分量图进行处理,标记出图像的荔枝果实,然后利用纹理统计法、颜色特征与果实不同部分面积比值相结合的方法进行果实变质的判断,最终实现了未成熟、成熟以及腐烂变质的荔枝果实的视觉智能判断,建立了荔枝果实品质辨识的智能系统。试验结果表明,辨识荔枝品质状态的正确率达93%。

Fruit Quality Detection Based on Machine Vision Technology when Picking Litchi

In order to judge fruit quality real timely, three conditions of litchi fruit, immature, mature and appearance rot after mature, were analyzed by using the fruit images of different growth periods in natural environment. The YCbCr color space model was selected and the exploratory analysis method was used to analyze and estimate the Cr component of litchi images of different parts, different illuminations and different growing periods, and the threshold ranges of Cr components of mature and immature litchi fruits was determined; for mature litchi, the fruit edge detection and Hough circle fitting processing were carried out on the Cr component diagram to mark litchi fruits. And then the texture statistics and the method that combining the color feature and area ratio of different parts of litchi fruit were used to judge litchi fruit deterioration. Finally, the vision intelligent judgment for the immature, mature and appearance rot of litchi fruit was realized and an intelligent system to identify litchi fruit quality was built. The test results show that the accuracy of identify litchi fruit quality is 93%.