Research Article

A Crack Identification Method for Concrete Structures Using Improved U-Net Convolutional Neural Networks

Table 2

Comparison of the measured crack widths.

Crack numberMeasured resultsACTIS measurementsRelative error

10.740.3552.7
21.100.4063.6
30.810.4149.4
40.350.2820.0
50.330.2427.3
60.270.2411.1
70.370.3116.2
80.430.3518.6
90.360.2725.0
100.430.2444.2
110.360.2530.6
120.480.3037.5
130.370.1948.6
140.420.3419.0
150.350.2334.3
160.580.3441.4
170.370.3116.2
180.560.3930.4
190.570.4619.3
200.450.3131.1