Research Article

Development of Deep Learning Model for the Recognition of Cracks on Concrete Surfaces

Table 1

Summary information of the database.

ParameterValue and description

Total number of images40,000
Image size (pixel)227227
Number of images with cracks20,000
Number of images without cracks20,000
Split distributionUniform, 50/50
Number of images in the training dataset20,000
Number of cracked images in the training dataset10,000
Number of no-cracked images in the training dataset10,000
Number of images in the testing dataset20,000
Number of cracked images in the testing dataset10,000
Number of no-cracked images in the testing dataset10,000