As semiconductor devices continue to shrink, novel materials (e.g. (Si)Ge, III/V) are being tested and incorporated to boost device performance. Such materials are difficult to grow on Si wafers without forming crystalline defects due to lattice mismatch. Such defects can decrease or compromise device performance. For this reason, non-destructive, high throughput and reliable analytical techniques are required. In this paper Electron Channeling Contrast Imaging (ECCI), large area mapping and defect detection using deep learning are combined in an analytical workflow for the characterization of the defectivity of “beyond Silicon” materials. Such a workflow addresses the requirements for large areas 10-4 cm2 with defect density down to 104 cm-2.