Abstract
|
Article Information:
Image Segmentation and Maturity Recognition Algorithm based on Color Features of Lingwu Long Jujube
Yutan Wang, Jiangming Kan, Wenbin Li and Chuandong Zhan
Corresponding Author: Jiangming Kan
Submitted: August 19, 2013
Accepted: August 27, 2013
Published: December 05, 2013 |
Abstract:
|
Fruits’ recognition under natural scenes is a key technology to intelligent automatic picking. In this study, an image segmentation method based on color difference fusion in RGB color space was proposed in order to implement image segmentation and recognition maturity intelligently according to Lingwu long jujubes’ color features under the complex environment. Firstly, the three-dimensional histograms of each color component which is widely used in color space currently are compared; and then the jujubes’ red area and non-red area was extracted respectively, thus, the whole target area is obtained by sum of those areas; then, watershed algorithm combined with mathematical morphology distance and gradient was utilized to overcome adhesion and occlusion phenomena; finally, the maturity level was recognized by the established recognition model of Lingwu long jujubes. The segmentation was tested through 100 sample set and 93.27% of precision rate was attained, so was correct estimating rate of maturity level recognition above 90%. The results indicate that a smaller average segmentation error probability is in this method, which is more efficient in the extraction and recognition of jujubes with red and green and the problem of segmentation and maturity level judgment of adhesive fruits is solved by the method as well.
Key words: Color difference, image segmentation, lingwu long jujubes, mature level, watershed transform, ,
|
Abstract
|
PDF
|
HTML |
|
Cite this Reference:
Yutan Wang, Jiangming Kan, Wenbin Li and Chuandong Zhan, . Image Segmentation and Maturity Recognition Algorithm based on Color Features of Lingwu Long Jujube. Advance Journal of Food Science and Technology, (12): 1625-1631.
|
|
|
|
|
ISSN (Online): 2042-4876
ISSN (Print): 2042-4868 |
|
Information |
|
|
|
Sales & Services |
|
|
|