This paper develops a method of knowledge-aided clutter suppression for subsurface radar imaging by crafting the power spectrum of the transmit waveform. The power spectrum is optimized to enhance the signal-to-clutter-ratio (SCR) by target matched illumination for suppressing transmit power in dominant clutter frequencies. This approach is extended to radar imaging for the enhancment of the overall 2-dimensional target reflectivity function. By modelling the received signal as the output of an LTI filter, the channel capacity can be maximized. An algorithm is developed using Taguchi techniques to optimized the spectrum and thus reduce the clutter in radar images. Particle swarm optimization (PSO) is extended to waveform design for comparison. The optimization objective function is the mutual information between the received signal and the target response conditioned upon a transmit waveform. The PSO and Taguchi algorithms are seen to converge quickly to a waveform that maximizes the SCR. The radar imagery is formed using the newly optimized waveforms to test the methodology by the coherent addition of simulated and distributed target and clutter responses. The SCR is used as a performance metric. The results show that transmitting a waveform tailored to the frequency response of the target and clutter improve the SCR and mutual information over a conventional linearly frequency modulated waveform.