基于图像传感技术的娃娃菜外观品质检测

目的 设计一套基于图像传感技术分析娃娃菜外观品质如尺寸、重量、瑕疵点等的检测方法.方法 搭建一套图像采集平台拍摄娃娃菜不同侧面,应用图像处理技术,分割出娃娃菜图像区域,并数字化其区域特征信息(包括:投影面积、尺寸、瑕疵点面积等).结果 建立合格娃娃菜的侧面投影面积与重量真实值间的线性关系,其相关系数为0.938,均方根...

Detection of external quality for baby cabbage by image sensing technology

Objective A detection method was proposed on the basis of image sensing technology to analyze the external indicators of baby cabbage, including the size, weight, defects.Methods An image acquisition platform was built to obtain images from different sides of baby cabbage, and the image processing technology was employed to segment image area and digitize its regional feature information (including: the projecting area, size, defected area, etc.). Results A linear model was built between the projecting area and the true weight for the qualified baby cabbages, with correlation coefficient of 0.938, as well as the root mean squared errors of 36.52 g. Compared with manual detection of surface defects, image recognition can reach the accuracy of 95 % for defects like rotted spots, cracks, etc. According to the grading technical standards of baby cabbage, two clustering algorithms were used to cluster baby cabbage samples into three grades based on a series of image characteristic descriptors. One is the K-medoid method reached the accuracy of 100 %, and another is the fuzzy Gath-Geva reached 96.67 %, and both of them got the satisfied the accuracy of clustering. Conclusion Results showed that the machine vision could be applied to automatic detection and classification of baby vegetables, which provided an effective method for online non-destructive detection of vegetable.