A Handheld Terminal Device Design For Concrete Pavement Disease Detection System

In order to improve the efficiency of detecting the concrete pavement, in this paper a handheld terminal embedded devices is designed for collecting pavement distress information. The veracity and rapidity of collecting pavement distress information from highway maintenance department is achieved by this system.First, the requirements for the equipment from the highway maintenance department and the types of concrete pavements are researched. After confirming their needs, the overall project is designed. The Samsung’s S3C2440processor is the core in this project, and the peripheral equipment include OV9650CMOS Camera, Liquid Crystal Display, touch screen and USB interface. In addition, transplant embedded Linux operating system as a handheld’s OS, then Uboot and Linux kernel are transplanted, and the Yaffs Root System is made. The OV9650drove by V4L2is in normal work. On this basis, Qtopia is transplanted as a beautiful desktop system, portable SQLite database is transplanted for managing the pictures and its’ disease information. Finally, the application software is developed by Qt.In order to realize the concrete pavement cracks disease is processed automatically, the automatic detection method for concrete pavement cracks is studied. The original image must be preprocessed so that the image is identified and processed easily by the computer. Preprocessing include gray image, filtering processing and image gray-scale stretching processing. Otsu method is used to segment the crack images; this method can avoid the misjudgment of the situation caused by the non-crack stripes, but the edge algorithm cannot. On the basis of the quantity difference of the background connected regions in the linear and reticular crack image, the paper presents the region labeling method to search the background connected regions divided by the crack, and realizes the crack classification. The three kinds of linear crack, the transverse, longitudinal and diagonal crack are subdivided by projection method. At last, the damaged area in the reticular crack image is calculated by finding its minimum enclosing rectangle. And the length of linear crack is measured after its skeleton is extracted by thinning algorithm, then the mean width of linear crack is calculated by its length and area.