Detection Model of Eggs Based on Improved BP Neural Network

Abstract:

In order to improve the precision in the detection of egg fresh degree, an improved BP neural network ensemble method was proposed herein, where the K-means clustering was applied to optimize better neural network individuals and Lagrange multiplier was used to not only compute the weight of neural network individuals, forecast egg fresh degree (Haff value). Based on the image of eggs acquired by the machine vision device, taking color characteristic parameter (H, S and I) of the central area of the egg as input and Haff value of the egg as output, a model was constructed. With experimental verification, it was calculated that the mean square error of Haff value was 1.3764 and the generalization ability of the network was high.

• Periodical: Applied Mechanics and Materials (Volumes 427-429)
• Pages: 564-566
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.427-429.564
• Online since: September 2013
• Authors: Jian Yang*, Chun Yan Xia, Zhan Wu Peng
• Keywords: BP Neural Network (BPNN), Egg Fresh Degree, K-Means Clustering, Nondestructive Testing (NDT)

References

[1] Q.H. Wang, Y.L. Ren and Y.X. Wen : Transactions of the Chinese Society for Agricultural Machinery Vol. 37 (2006) , p.104.

[2] L.M. Zheng, X. Yang , G.Y. Xu and H.Y. Guo: Transactions of the Chinese Society of Agricultural Engineering Vol. 10(2009), p.335.

[3] D.F. Zhang: MATLAB Neural Network Application Design. (China Machine Press，China 2012).

[4] K. Li and H.K. Huang: Journal of Computer Research and Development Vol. 42(2005), p.594.

[5] L. Breiman: Machine Learning Vol. 24(1996), p.123.

[6] K. Li and H.K. Huang: Journal of Computer Research and Development Vol. 42(2005), p.594.