International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 105 - Number 18 |
Year of Publication: 2014 |
Authors: Xuming Chen, Simon X. Yang |
10.5120/18475-9425 |
Xuming Chen, Simon X. Yang . Configurable Solution for Object Recognition and Localization based on Features in YUV Color Space. International Journal of Computer Applications. 105, 18 ( November 2014), 6-12. DOI=10.5120/18475-9425
In this paper, a practicable robot vision solution is presented for multiple kinds of object recognition and localization. The proposed solution generalizes and extends simplicity driven algorithms for recognition and localization of fruits under harvesting environment. In the proposed solution, fruits with shape of quasi round such as oranges, apricots and apples on the tree are segmented from complicated real background with one dimensional V computing in YUV. Further object recognition algorithms are applied to recognize and compute each qualified fruit's localization. For a study of generality, the original solution is extended from fruit recognition to logistics and production industry. One is for recognition and localization of mixed kinds of cookies on conveyor belt in packaging line. Another one is for recognition of parts in warehouse or production place with a synthetic color labeling methods. V-U difference is used to enhance the difference among different kinds of objects. By a certain feature of one-dimension, multiple kinds of objects are distinguished from each other. Based on reconfigurable modeling, the recognition of multiple kinds of objects is realized on a single system.