Skip to Main Content
Least Squares Support Vector Machines could satisfactorily describes the non-linear relationships between the image information and the 3D information. It doesn't need to confirm internal and external parameters of the camera. The kernel function parameter and penalty parameter is a pivotal factor which decides performance of LS-SVM. Most users select parameters for an LS-SVM by rule of thumb, which frequently fail to generate the optimal approaching effect for the function. In order to get optimal parameters automatically, an adaptive genetic algorithm is introduced to the LS-SVM algorithm,which automatically adjusts the parameters for LS-SVM. The experimental results show that X, Y axis error values of AGA-LS-SVM is smaller than LS-SVM by 2~3 times, and Z axis error values of AGA-LS-SVM is smaller than LS-SVM by 10 times. The validity of improving the calibration accuracy is verified by experimental results.