Influence of Image Gray Process on the Detection of Pavement Cracks
The longitudinal crack,transverse crack and alligator crack was considered as research objectives,the crack automatic detection was performed by YOLOv7 algorithm in this study.The image dataset was conducted by grey operation to obtain the colorful image set and the grey image set.In terms to the above two image sets,the in-fluence of algorithm parameters of the iteration time,confidence threshold and the intersection to union ratio threshold(IOU)on the accuracy and performance was analyzed to propose the optimal parameter combinations.The field measured images were utilized as input for the detection model.Based on the color image set and the gray image set,the pavement crack was detected.The influence of the image grey operation on the detection accuracy was stud-ied.The result indicates that all the effects of parameters on the prediction accuracy showed an obvious increase and followed by a decrease trend.The color image could capture more sufficient information of the pavement characteris-tics compared with gray image.The crack disease identification based on color image set achieved greater accuracy and more efficient performance.The research can provide an inference for the accuracy improvement of pavement crack automatic detection.
road engineeringasphalt pavementcrack diseasedetection accuracyimage grey process
国家科技期刊平台
NSTL
万方数据
