some of the GPR electronics fellows could probably give you a more detailed answer, but in a lame term perspective, I believe that you have an obvious problem here.
GPR antennas are and should be UWB (ultra wide band) devices, so unlike in the communication world, you want your antenna to have as wide bandwidth as possible to improve the data. From that perspective feeding the antenna with the generator or a simple dipole may not be the best choice for the antenna design.
Narrow banded antennas (in GPR world) are bad, because they have a tendency to be very "ringy" (horizontal banding in the data). That means that most of the energy is used for "useless" repeating of a single frequency (and harmonics related). The penetration will be limited and any responses (reflections) will be overwhelmed with the ringing, so you will have data that is terrible to interpret.
One way of evaluating the quality of a GPR antenna is by looking at the bandwidth.
Example commercial 100MHz antenna
I hope this helps you a little bit in your work.
Thank you. So you think my antenna is the issue. In that case can I use a 200 MHz centre frequency antenna with 100 - 300 MHz bandwidth bowtie antenna for the same 10 ns width pulse signal??