Get the Cannon Box near the Stamp and sink down to the bottom of the area. You will find a cracked wall here. Destroy it with a Cannon Box shot and you will be transported to a new area. Get the Green Star here, and a bunch of loot to boot!
The development of bridge crack-detection methods has been relatively slow. Traditional manual detection is not only time-consuming and laborious but also has many unsafe factors. The bridge inspection vehicle is a special vehicle that can provide a working platform for bridge inspection personnel during the inspection process and is equipped with bridge inspection instruments for flow inspection and/or maintenance operations. However, its utility is limited by its high production cost and complex manufacturing process.
Non-destructive testing technology has been widely used in the field of bridge crack detection. Common non-destructive testing methods include optical fiber sensing , ultrasonic detection , and acoustic emission detection . However, these non-destructive methods have some limitations. Optical fiber sensing technology requires the laying of optical fibers, which is expensive. Acoustic detection technology is only suitable for detecting cracks in a single direction of a bridge deck with a small detection range. Acoustic emission detection technology can only detect cracks that are being generated at present and cannot detect cracks that have previously formed. Therefore, the high detection cost, limited working conditions, and inefficient detection speed limit the traditional detection methods based on manual detection or instrument information characteristic analysis, and it is important to devise a new technical means to carry out real-time and efficient bridge crack detection.
The second type is a regression-based object detection model represented by a single-shot multi-box detector (SSD)  and unified real-time object detection (YOLO) . Compared to object detection models based on candidate regions, regression-based object detection models have a faster detection speed.
Author response: Special thanks to you for your good comments. This paper designs a bridge crack detection network based on the deep learning theory. The work of this paper is only a part of the bridge crack detection project. The advantage of the detection network based on computer vision is its learnability. The network can learn the relevant knowledge of crack detection from the data samples and can detect the bridge crack accurately and quickly by relying on the mighty computing power of the convolution neural network. Compared with other traditional detection technologies, crack detection methods mainly rely on artificial vision inspection or instrument signal characteristic analysis. The detection accuracy and speed lag behind the detection network designed in this paper. However, bridge crack detection is a complex engineering project. The whole project includes many hardware, software, and corresponding manual operations. The design of bridge crack in this paper is a systematic engineering project, so it is impossible to compare with other detection technologies from the overall perspective of bridge crack detection engineering.
Author response: Special thanks to you for your good comments. Network parameters and training mode play a vital role in the final performance of the network. Combined with the characteristics of bridge crack detection, we show more detailed information on the network training process in the article and comprehensively analyze the network's performance in different gradient descent modes and different iterative stages.
Reviewer#2, Concern # 5: References and the text are not consistent: is not work of Zhao et al. Lu et al.'s work  is not about bridge crack detection as asserted by the authors. Feng et al.'s wok  is not about bridge crack detection. However, on can stress that they are correlated problems. Just align the previous work and your statements.
Author response: We gratefully appreciate for your valuable suggestions. Bridge cracks and pavement cracks are both concrete surface cracks. From the perspective of network engineering applications, we expect that the network designed in this paper is not limited to bridge crack detection, so we refer to many similar cracks detection research. 2b1af7f3a8