Ground Penetrating Radar (GPR) is capable to detect plastic anti-personnel landmines as well as other subsurface targets. In order to reduce false alarms, an option of automatic landmine discrimination from neutral mine-like targets would be very useful. This paper presents a possibility for such discrimination and analyzes it experimentally. We investigate time-frequency features of an ultra-wideband (UWB) target response for the discrimination between buried landmines and other objects. The discrimination method includes the extraction of an early-time target impulse response, its time-frequency transformation and the extraction of time-frequency features based on singular value decomposition (SVD) of the transformed image. In order to take into account changes in the UWB target signals, the experimental conditions are completely controlled to focus on the behaviour of the target{\textquoteright}s response with respect to its depth and the horizontal position of the GPR above it. The dependence of the features on the GPR bandwidth is analyzed as well. The Mahalanobis distance is used as a criterion for optimal discrimination. The obtained results define the best features and conditions when landmine discrimination is successful. For comparison, the discriminant power of the proposed features has been tested on a dataset, acquired during a field campaign in Angola.