Multi-color space threshold segmentation and self-learning k-NN algorithm for surge test EUT status identification

Abstract

The identification of targets varies in different surge tests. A multi-color space threshold segmentation and self-learning k-nearest neighbor algorithm (k-NN) for equipment under test status identification was proposed after using feature matching to identify equipment status had to train new patterns every time before testing. First, color space (L*a*b*, hue saturation lightness (HSL), hue saturation value (HSV)) to segment was selected according to the high luminance points ratio and white luminance points ratio of the image. Second, the unknown class sample S_r was classified by the k-NN algorithm with training set T_z according to the feature vector, which was formed from number of pixels, eccentricity ratio, compactness ratio, and Euler’s numbers. Last, while the classification confidence coefficient equaled k, made S_r as one sample of pre-training set T_z′. The training set T_z increased to T_z+1 by T_z′ if T_z′ was saturated. In nine series of illuminant, indicator light, screen, and disturbances samples (a total of 21600 frames), the algorithm had a 98.65% identification accuracy, also selected five groups of samples to enlarge the training set from T₀ to T₅ by itself.

Keywords

multi-color space / k-nearest neighbor algorithm (k-NN) / self-learning / surge test

References

[1] Williams T, Flood J. Guidance note addresses surge test problems. Compliance Engineering, 2002, 19(11): 47–49

[2] da Silva D A, da Costa E C M, De Franco J L, Reliability of directly-molded polymer surge arresters: Degradation by immersion test versus electrical performance. International Journal of Electrical Power & Energy Systems, 2013, 53: 488–498

[3] Xu Y, Maeno K, Nagahara H, Camera array calibration for light field acquisition. Frontiers of Computer Science, 2015, 9(5): 691–702

[4] Wang X, Wu X, Zhang X. Multi-resolution texture synthesis from turntable image sequences. Frontiers of Mechanical Engineering, 2012, 7(1): 72–80

[5] Amato G, Falchi F, Gennaro C. Geometric consistency checks for kNN based image classification relying on local features. In: Proceedings of 4th International Conference on Similarity Search and Applications, SISAP 2011. New York: ACM, 2011, 81–88

[6] Huang J, Ding J, Liu J, Citation-kNN algorithm based on locally-weighting. Journal of Electronics & Information Technology, 2013, 35(3): 627–632 (in Chinese)

[7] Du S, Liu Z. A comparative study of different color spaces in computer-vision-based flame detection. Multimedia Tools and Applications, 2015, 1–20

[8] Soleimanizadeh S, Mohamad D, Saba T, Recognition of partially occluded objects based on the three different color spaces (RGB, YCbCr, HSV). 3D Research, 2015, 6(3): 1–10

[9] Chernov V, Alander J, Bochko V. Integer-based accurate conversion between RGB and HSV color spaces. Computers & Electrical Engineering, 2015, 46: 328–337

[10] Peng Z, Chang F, Dong W. Vibe motion target detection algorithm based on lab color space. Communications in Computer and Information Science, 2015, 525: 45–54

[11] Kaarna A, Liu W, Kälviäinen H. Development of color density concept with color difference formulas in respect to human vision system. Frontiers of Electrical and Electronic Engineering in China, 2011, 6(2): 381–387

[12] Chen X, Liu G, Huang J. Method of image acquisition and curved surface eliminating for flexible printed circuit. Journal of Electronic Measurement and Instrumentation, 2015, 29(6): 895–900 (in Chinese)

[13] Huang J, Liu G, Wang X. Multi-processor scheduling of vision detecting system for filter surface defect. China Measurement & Test, 2015, 41: 90–93(in Chinese)

Acknowledgements

The authors wish to acknowledge the financial support from the Science and Technology Plan Projects of Guangzhou, China (Grant No. 201504010037).

RIGHTS & PERMISSIONS

2016 Higher Education Press and Springer-Verlag Berlin Heidelberg