Road crack detection and measurement is one of the most important challenges for driving safety. Although the problem has been studied for a long time, few researchers have been investigating the road crack real-world measurements. Here, we develop an intelligence approach that detects the road crack then applies image processing techniques for the segmentation and real-world measurements. Our method demonstrates to be able to detect and measure the crack real-world area. The present intelligence approach including improved architecture CNN-YOLOv5 (Convolutional Neural Network-You Only Look Once version5), YOLO-crack were proposed to improve the ability of the network to detect multiple cracks. On the one hand, YOLO-crack realizes cracks detection on images. In CNN-YOLOv5, a hybrid module is proposed, which combines CNN and YOLOv5 to extract sparse the multi-scale features expressed can avoid the loss of local information caused by dilated convolution.