Ultrasonic imaging techniques are typically deployed using plane wave beamforming such as delay-and-sum or time migration methods. In complex parts such as stiffened composite panels, wave refraction and dispersion effects complicate the beamforming process. In these cases, the complexity of the propagating waves can be leveraged for imaging damage locations by using data-driven Matched Field Processing (MFP) techniques. These methods are effectively used in underwater acoustics but much less employed in NDE. This work presents active imaging techniques within an MFP framework using a sparse array of ultrasonic piezoelectric transducers to image simulated damage in a stiffened composite panel. In particular, the coherent white noise constraint (C-WNC) algorithm achieves both high resolution and high contrast in imaging wave scatterers in waveguide structures. Techniques of normalized cross power spectrum, dominant source projection, and adaptive beamforming are demonstrated to obtain high imaging performance in these complex structures.