Pixel-Level Recognition of Pavement Distresses Based on U-Net

<table class="table-group" id="tab6"><tr><td><table class="table"><tr><td class="thead-hr" colspan="2"><hr/></td></tr><tr class="thead"><td class="align_left">Distress</td><td class="align_center">IoU</td></tr><tr><td class="thead-hr" colspan="2"><hr/></td></tr><tr><td class="align_left">No distress</td><td class="align_center">0.99059</td></tr><tr><td class="align_left">Patch</td><td class="align_center">0.73778</td></tr><tr><td class="align_left">Pothole</td><td class="align_center">0.28436</td></tr><tr><td class="align_left">Transverse crack</td><td class="align_center">0.21429</td></tr><tr><td class="align_left">Block crack</td><td class="align_center">0.19445</td></tr><tr><td class="align_left">Alligator crack</td><td class="align_center">0.23038</td></tr><tr><td class="align_left">Longitudinal crack</td><td class="align_center">0.12581</td></tr><tr><td class="align_left">Average (MIoU)</td><td class="align_center">0.3968</td></tr><tr class="table-tr"><td colspan="2"><hr class="tbody-hr"/></td></tr></table></td></tr></table>

<div>IoU of different labels using ResNet-34.</div>

Advances in Materials Science and Engineering

tab6

Table 6

Table 6: Pixel-Level Recognition of Pavement Distresses Based on U-Net 