Online variety recognition of auto rack girders based on combination of Fuzzy ART neural network with D-S evidence theory

Wang, Hua; Gao, Jingang; Zhang, Shuang

doi:10.1504/ijmic.2009.029264

Cited by 2 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is not appropriate for typical products in mass production on-line detection. [1,2,3] On account of the auto rack girder holes number, diameter and position precisely and quickly detection, it is presented that high speed camera distortion correction based on radial base function (REF) neural network, at the same time corresponding relationship between REF neural network and auto rack girder pilot holes is built. It is presented that a pair of source of light is added into auto rack girder pilot holes detection system.…”

Section: Introductionmentioning

confidence: 99%

Study of auto rack girder detection system

Wang

Gao

Zhang

2010

2010 International Conference on Computer, Mechatronics, Control and Electronic Engineering

Self Cite

View full text Add to dashboard Cite

On account of out-of-date and low efficiency detection mode for large dimension auto rack girder, it is presented that is the Machine Vision on-line detection system for auto rack girder Based on RBF and linear structured light.It is corrected according to the error of RBF neural network for camera lens. Taken advantage of RBF characteristic that approximating high order input and output nonlinear system, camera image error correction high order distortion model is constructed and the precision of camera detection is raised. This correction method is used in China FAW GroupCorporation for auto rack girder detection system. Automatic non-contact detection is firstly realized in China for auto rack girder holes diameter, position and number. After tested at production field, the precision of detection diameter is ± O.lmm, the precision of detection position is ±O.3mm, and the detection time is less than 1.5 minutes. It is proved in practice that the method of linear array CCD camera lens distortion correction based on RBF neural network is feasible.

show abstract

Section: Introductionmentioning

confidence: 99%