Gabor wavelet application in image processing

hello, I used the wavelet of Gabor for the extraction of image texture features. I calculate the characteristics of Gabor in 5 scales and 6 orientations and with the following parameters: minWaveLength = 3; mult = 1.7; sigmaOnf = 0.65; dThetaOnSigma = 1.3; k = 0; polarity =0. and I got filters like these:

which is completely different from the results obtained in the links I liked: http://www.peterkovesi.com/matlabfns/PhaseCongruency/Docs/convexpl.html

For the angular component of filter (within 6 orientations) I got:

For the radialcomponent of filter (within 5 orientations) I got:

I do not know if I hurt chosen filter parameters or if there is another problem because the image obtained after the convolution does not visualize the edges (black image). and for the classification results of the texture FEATURES of the images, I obtained a low accuracy. Any idea about these results ??

Matlab code: http://www.mathworks.com/matlabcentral/mlc-downloads/downloads/submissions/55096/versions/1/previews/Soil%2520Detection_Upload/Soil%2520Detection_Code/gaborWavelet.m/index.html