Research On Crack Detection Method Of Concrete Bridge Based On Infrared Image Processing

Bridges play an important and irreplaceable role in China’s transportation and economic development.In the process of using concrete bridges,cracks caused by environmental temperature,aging of building materials and other factors are an important disease of bridges.They not only affect the normal use of bridges,but also cause other diseases,seriously reducing the service life and safety.Therefore,it is very necessary to adopt efficient method to detect the cracks of the bridge regularly.At present,the detection of bridge cracks still mostly depends on manual work,which is dangerous,laborious,timeconsuming and costly,and cannot meet the demand of bridge defect detection in China.With the development of computer technology and digital image processing technology,crack detection technology based on image processing is getting more attention.Due to the complex geographical and natural environment and concrete surface roughness,the collected concrete images have problems such as spalling,stains,scratches,and uneven illumination,which affect the detection accuracy.Most of the crack detection algorithms based on image processing taken by ordinary camera cannot address these problems effectively and simultaneously.Compared with ordinary images,infrared image detection greatly reduces the impact of these problems.Any object produces thermal radiation.The intensity of infrared radiation is related to the material type,structural characteristics and temperature of the measured object.In the application of crack detection of concrete bridges,the infrared thermal radiation distribution between the crack area and the surrounding concrete surface area is different and can be converted into the form of infrared image,and then the detection of crack defects can be realized through image processing to determine the shape and location of cracks.This paper proposed a real-time crack detection method for bridge concrete based on infrared images.And,a set of embedded system for real-time crack detection is designed,which is suitable for carrying on the portable equipment such as unmanned aerial vehicles(UAVs)and handheld equipment.The system uses the TC388 infrared sensor module to collect the image of the concrete bridge surface,and uses the ARM CortexA8 embedded processor to process,display and detect the cracks in the infrared image,and finally store the crack image data.For the case where there are stains,peeling,scratches,and uneven illumination on the surface of concrete bridges,and where it is difficult to accurately detect complete cracks,this paper proposes a new method to connect the breaks in cracks by adaptive morphological dilation based on crack direction.Most of the existing crack image detection methods attempt to achieve high detection accuracy by increasing the algorithm complexity but sacrifice realtime detection efficiency.A multiple filtering method based on a few adaptive feature thresholds is proposed to filter non-cracks and obtain a clear crack image by analyzing the morphological characteristic differences between real cracks and noise and pseudo-cracks.Finally,using the actual infrared images of the concrete surface for experiments.The results show that the proposed method in this paper can realize fast and accurate detection of bridge cracks,and can effectively improve the integrity of cracks,remove the interference of different noises and pseudo-cracks,and the accuracy of crack detection is 96.99%.This method does not require modeling and has a fast detection speed.Combined with the bridge crack embedded detection system designed in this paper,it is suitable for practical engineering applications.