Affiliation:
1. Dalian Polytechnic University
2. Jilin University
Abstract
In order to realize the intelligent recognition for the maturity grade of fresh corn ear, intelligent inspection system was studied based on computer vision, which could automatically complete the collection and handling of ear graphic and the recognition of maturity of the corn ear. Based on the study, a kind of intelligent recognition method was put forward under the graphic of certain frequency domain. An energy chain was established, and the characters of energy spectrum was extracted through the two-dimensional inverse discrete Fourier transformation on the graphics collected. With the above characters, a probabilistic neural network was developed, the accuracy rate of the recognition method could be 96.7%.
Publisher
Trans Tech Publications, Ltd.
Reference6 articles.
1. PANIGRAHI S, MISRA M K, BERN C, Background Segmentation and Dimensional Measurment of Corn Germplasm, Transactions of the ASAE, vol. 38, Jan. 1995, pp.291-297.
2. LIU J and PAULSEN M R, Corn Whiteness Measurement and Classfication Using Machine Vision, Transactions of the ASAE, vol. 43, 2000, pp.757-763.
3. Shuqin Yang, Jifeng Ning and Dongjian He, Identification of corn breeds by BP neural network, Journal of Northwest A&F University, vol. 32(supp), 2004, pp.162-164.
4. Yunlong Zhou, Zhenru Fan and Yaolei Su, Identification method of flow regine based on images texture features by fourier transform and probablistic neural network in gas-solid fluidized bed's, Control and Instruments In Chemical Industry, vol. 36, 2009, pp.38-42.
5. KUO LIANG, CHUNG and WAN YUCHEN, Fast adaptive PNN-based thresholding algorithms, Pattern Recognition, vol 36, Dec. 2003, pp.2793-2804.