Laser-induced ultrasound in metals challenges the standard pulsed photoacoustic model because the thermal confinement approximation does not hold for materials with optical absorption \(\varvec{\ge 10}^{\varvec{6}}\text {m}^{\varvec{-1}}\) . Recently, we heuristically removed this approximation a posteriori from the one-dimensional (1D) solutions by modeling heat-flux generation and propagation during a laser pulse [1]. Here, we evaluate whether these 1D solutions are sufficient for analyzing two-dimensional B-scan images of subsurface cavities and microdefects. To this end, we fabricated a test structure with cavities by joining two metal plates. We acquired B-scan data and conducted a theoretical study of heat-flux effects, which was experimentally validated. Using the 1D model directly on the measured signals, we generated images and metrics and quantified the accuracy. We found that heat flux reduces reflection amplitudes while the time of flight (TOF) remains unchanged. The TOF estimation based on the model significantly improves image definition and contrast, resulting in a noise-free image. The 1D model locates and dimensions hidden cavities with a discrepancy of \(\varvec{\simeq 3.5-5.0\%}\) (100–300 \(\varvec{\mu m}\) ) relative to vernier caliper measurements, consistent with a pixel size of 63.5 \(\varvec{\mu }\) m; comparable performance is achieved for the clamped joint. These results support minimally supervised LIUS metrology based on 1D model images in metals
