Transverse crack pattern significantly affects performance of continuously reinforced concrete pavement (CRCP). However, field and modeled crack spacing are based on visual surveys of fully developed transverse cracks visible on the slab surface. Because of the difficulties in detecting incipient cracks in concrete slabs, the effect that such cracks may have on CRCP performance is not currently considered. To address this issue, this paper presents a non-destructive evaluation methodology based on an ultrasonic linear-array system for detection of incipient cracks in CRCP slabs. Two ultrasonic signal-processing techniques were used. First, the Hilbert Transform Indicator was used to identify the presence of damage in the concrete slab indicating potential crack locations. Then ultrasonic image reconstructions of these locations were used to further evaluate crack presence. The methodology was applied to four sections of an experimental short CRCP composed of 50-m long slabs – short in comparison to traditional CRCP. The locations of 58 potential incipient cracks were identified updating the short CRCP crack spacing; 10 of these incipient cracks were confirmed as surface-visible cracks in later visual surveys. Additionally, the methodology shows potential to detect undesired crack patterns such as cluster and Y-cracking before the cracks emerge on the slab surface.