ON-LINE DETECTING SIZE AND COLOR OF FRUIT BY FUSING INFORMATION FROM IMAGES OF THREE COLOR CAMERA SYSTEMS
On the common systems, the fruits placed on rollers are rotating while moving, they are observed from above by one camera. In this case, the parts of the fruit near the points where the rotation axis crosses its surface (defined as rotational poles) are not observed. Most researchers did not consider how to manage several images representing the whole surface of the fruit, and each image was treated separately and that the fruit was classified according to the worse result of the set of representative images. Machine vision systems which based 3 color cameras are presented in this article regarding the online detection of size and color of fruits. Nine images covering the whole surface of an apple is got at three continuous positions by the system. Solutions of processing the sequential image’s results continuously and saving them into database promptly were provided. In order to fusing information of the nine images, determination of size was properly solved by a multi-linear regression method based on nine apple images’ longitudinal radius and lateral radius, and the correlation coefficient between sorting machine and manual is 0.919, 0.896 for the training set and test set. HSI (hue-saturation-intensity) of nine images was used for apple color discrimination and the hue field in 0o~80o was divided into 8 equal intervals. After counting the pixel in each interval, the total divided by 100 was treated as the apple color feature. Then 8 color features were got. PCA and ANN were used to analysis the 8 color features. There is a little overlapped in the three-dimensional space results of PCA. An ANN was used to build the relationship between 8 color characters and 4 apple classes with classification accuracy for the training/test set 88%/85.6%.
KeywordsPrincipal Component Analysis Classification Accuracy Color Feature Fuse Information Rotational Polis
- Li Q. Z., Wang M. H. (1999) Study on High-Speed Apple Surface Defect Segment Algorithm Based on Computer Vision. Proceedings of 99 International Conference on Agricultural Engineering, Beijing, China, December.Google Scholar
- Miller W M (1995). Optical defect analysis of Florida citrus. Applied Engineering in Agriculture, ASAE, 11(6), 855–860Google Scholar
- Penman DW(2002). Determination of stem and calyx location on apples using automatic visual inspection. Computers and Electronics in Agriculture, 33(2002), 7–18Google Scholar
- Tao, Y., & Wen, Z. (1999). An adaptive spherical image transform for high-speed fruit defect detection. Transactions of the ASAE, 42(1), 241–246.Google Scholar