We propose a disruptive approach based on μLaue diffraction to study the nature and the distribution statistics of defects in semiconductor materials. The fine structure of the g peaks of Laue patterns are modified by the presence of defects according to the g.b criterion used in e- microscopy to determine the b Burgers’s vectors of dislocations. Here we propose to finely analyze and map the shape of the Laue peaks (corresponding to hkl planes) under the polychromatic beam footprint. The sorting of the peak’s distortions/elongations will provide the nature of the defects (extinction criteria) as well as the density of the defects (amplitude/intensity of the broadening). The collection of hkl peaks images obtained during mapping will be processed by AI methods (variational auto encoders, VAE) to get a fast and concise analysis. This will be compared to simulations of Laue patterns for simple defects (screw/edge components). This method will be demonstrated on several GaN materials.