Maximizing the detection of thermal imprints in civil engineering composites via numerical and thermographic results pre-processed by a groundbreaking mathematical approach