Automated Pavement Distress Collection Techniques (2004)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

28 This chapter presents the various pavement condition data management procedures and hardware used by the vendors and agencies. Although one would like to address the details of computing platforms, types of storage devices, and vari- ous programs used to manage the myriads of data collected in a typical process, questionnaire results do not support such detailed discussion. It is clear that some data management issues are more logistical than hardware related (e.g., the linking of time and location on data files as they are generated in the field). Efficient data management has become an ever-increas- ing concern as more and more data are collected with a sin- gle pass of data collection vehicles. Now that many of those vehicles can collect images (both pavement and other road- way), roughness, rut-depth, and joint-faulting data with one pass over a roadway, the tools for managing all the data at once also are changing. As would be expected, most users, both agency and vendor, begin by using data management systems incorporated into the data collection vehicle by the equipment manufacturer. However, the investment in a fully equipped vehicle is substantial; therefore, vehicles generally cannot be replaced each time a significant improvement in data management hardware or software takes place. The result is that for most systems an almost constant updating process takes place. In view of the very costly collection and processing pro- cedures encountered with pavement condition data, one uni- versally addressed issue is that of data archiving. Virtually every vendor has a rigorous archiving process before data are delivered to the customer, and virtually every agency has a similar process once data are delivered either from a vendor or from agency data collectors. The result is that most data are archived at least twice. A widespread companion prob- lem is what to do with all that data. Although some agencies reported on special efforts to solve the problem, the approach by most appears to be to acquire larger storage areas. The problem was especially severe when most images were videotaped. That part of the problem has been greatly allevi- ated by the transition to digital images. Certainly, the rapid advances in electronic data storage capabilities over the recent past, which are expected to continue, will do much to reduce the future bulk, if not the lines of data stored. Because almost all of the data management software and handling procedures are proprietary, it is very difficult to characterize the industry as a whole. However, it is safe to say that the demands of the data collection process are so intensive that the latest and highest speed processors (now in the 2- to 4-GHz range) and the largest data storage devices (now upward to several terabytes) available will be featured on the newest equipment produced by the various vendors. On-board data management is often handled on computers classified as “industry hardened”; that is, they are designed for rough service such as that encountered in moving vehi- cles. There are certain data collection and processing activi- ties with data management requirements that have much in common across the various vendors and among the various agencies, such that at least the general approaches can be outlined as discussed in the following paragraphs. There is some need for the development of data standards for pavement distress data. Although numerous data formats and handling procedures are in use in the realm of pavement data collection, nothing has been standardized. This means that essentially every vendor can use different formats and procedures and that customers and others can be at a loss as to how a given data management system works. A primary reason for this is that individual vendors or agencies have developed nearly all systems; that is, there has been little cen- tralized effort to build systems that work for all. Standards would offer another tool that agencies and others could use in the development of data procurement contracts. It would be much easier to specify that a data management system meet a certain standard than as is now done with spelling out specific formats, equipment, etc. The general term applied to information about data is “metadata,” or data about data. Simply put, metadata are the background information that describes the content, quality, condition, and other appropriate characteristics of the data. The National States Geographic Information Council has de- veloped a metadata primer or “cookbook” approach to the creation of metadata (57). Although the primer specifically addresses GIS, the concepts are general and the primer offers the following description of the value and uses of metadata standards: Metadata serves many important purposes including data brows- ing, data transfer, and data documentation. Metadata can be organized into several levels ranging from a simple listing of basic information about available data to detailed documentation about an individual data set. At a fun- damental level, metadata may support the creation of an inven- CHAPTER FOUR DATA MANAGEMENT PROCEDURES

29 tory of the data holdings of a state or local government agency. Metadata [are] also important in the creation of a spatial data clearinghouse, where potential users can search to find the data they need for their intended application. At a more detailed level, metadata may be considered as insurance. Metadata insures that potential data users can make an informed decision about whether data are appropriate for the intended use. Metadata also insures that the data holdings of an agency are well docu- mented and that agencies are not vulnerable to losing all the knowledge about their data when key employees retire or accept other jobs (57 ). The issue of metadata standards for pavement condition data is in need of a critical evaluation and automated pave- ment data collection and processing efforts would benefit greatly from the application of these concepts. The specifics of data management were requested in the project questionnaire. The responses to that question are given in Tables B1 through B4 for surface distress, ride quality, rutting, and joint faulting, respectively. The great variety of responses reflected in those tables makes it very difficult to identify any consensus procedures, such variety highlighting the need for standardization of data management systems. IMAGE-RELATED DATA MANAGEMENT Image data management depends greatly on the means of image capture. Where the principal media are videotapes, those tapes typically are delivered to the user after archiving by the data collector (contractor or agency). Images are typ- ically stamped with the date and time, as well as the selected means of location reference. Alternatively, a companion data file is stamped with tape linkages so that time, date, and location-reference information can be integrated. Those com- panion files typically are temporarily stored on the hard drive of a computer in the data collection van for later archiving and removal to the user’s media (usually tape or removable hard drives). Depending on the specific data processing arrangement, there may or may not be an intermediate dis- tress identification and classification step (sometimes with an index calculation) before delivery to the user. As described in chapter three, some processes require digitalization of the video images for distress data reduction purposes. The final step usually involves the copying or installing of the trans- ported media or processed results to a workstation or com- puter for the users’ purposes. In the few cases in which users are still using film, images are handled in much the same way as for those results that are videotaped. Digital images are managed somewhat differently. They are typically temporarily stored on a hard drive in the data collection van for later archiving and removal to the user’s media. Again, image files typically will be integrated with files containing date and time stamps, as well as location- reference information. Files are substantial, requiring large data storage devices—typically hard drives for the day’s work—and then downloading to portable hard drives for transfer to either a processing workstation or directly to the end user. Some agencies also receive the data on CDs, DVDs, and high-capacity zip discs. As with videotapes, there may be an intermediate distress identification and classifica- tion step before the files are delivered to the user. Although most vendors provide little detail on data man- agement, the ICC describes its image workstation software on the company website (25). The software was designed to manage digital image data to expedite the distress rating process and to maintain rating data. Images from multiple cameras can be synchronized. Then, users can categorize, measure, rate, and save the distress information. That infor- mation can be printed and exported in several formats. The use of Microsoft SQL 7.0 technology makes the application network compatible while it also interfaces with Adobe Photo- shop 6.0 for special image processing. Other details of the application of the ICC and other software were described in chapter three. Metadata standards would be helpful in the management of image data. The digital imaging tutorial (24) makes the case for metadata standards in that area. The TRB draft circular (20) offers some comments on the management of pavement distress data, especially in the area of image compression. Without compression, the stor- age need for 1 km (0.6 mi) of pavement imaging at 4 m (13 ft) wide is approximately one GB at 2,048 pixels per lane or four GB for 4,096 pixels. Therefore, compression is widely used for image archiving and data management. The predominant compression method in use is that of JPEG, an imaging industry standard-setting body. Normally, there is some loss of information during compression, such that an original raw image will not be fully restored from the compressed image. However, a new compression standard labeled JPEG 2000 has been designed to overcome some of the loss, because a much greater degree of compression will result in a restored image quality similar to that for traditional JPEG. Two exam- ples are given in the TRB draft circular. The first example (Figure 15) is a restored JPEG image compressed 6:1. The 4,096 pixels (transversely), 4 meters wide FIGURE 15 A 4,096-pixel resolution image in JPEG format (line scan digital camera) (20).

image size is approximately 1.4 MB. The second example (Figure 16) is of the same pavement surface as in Figure 15, but compressed with JPEG 2000. The size of the image in Figure 16 is approximately 400 KB, with a compression ratio of approximately 20:1. The image quality and definition of both figures are comparable. However, it must be recognized that encoding and decoding JPEG 2000 images will take more computing power. SENSOR-RELATED DATA MANAGEMENT As with digital image files, those for sensor-related data are large. Data collection vehicles fire lasers or other sensors at high speeds, collecting myriads of data every few millimeters along and across the roadway. Several sensors will be devoted to the longitudinal profile, whereas as many as 37 may be devoted to transverse profile measurement, with several giga- bytes of data captured in just a few kilometers of roadway. Again, fast processors, sometimes working in tandem, and large data storage devices are required just to capture and process profile data. Sensor data also must be integrated with date, time, and location-reference information before they are useful to the user. All of this calls for special software, an example of which is provided by ICC in its Windows-based profiling software (25). This program allows the user to collect, store, process, and graph profile data from a single Windows application. The program manages data files while building, saving, and using real-time viewing applications and reports. In addition, it has built-in diagnostics for checking and calibrating the profiler’s sensors. All of the information is available in real time for printout or exporting to ASCII text, Excel, and other files. Although other vendors are less specific in the descriptions of their software, requirements of the various systems dictate 30 similar applications. Metadata standards, as mentioned, would be of great benefit in the management of the data. SUMMARY The storage and management of pavement condition data and images are common problems for both vendors and users of the data. Although the great volumes of data produced have overtaxed storage capabilities in the past, the data storage industry appears to be solving many problems with the intro- duction of ever-greater storage capacity devices. Other data management problems are being alleviated with the periodic introduction of increasingly faster processors, a trend that appears to have no end in sight. There may be a data storage problem at some agencies with regard to how to handle the large volumes of historical data, especially if earlier videotape systems were extensively employed. There appears to be a critical need for the development of metadata standards for automated pavement data collection and processing systems. 4,096 pixels (transversely), 4 meters wide FIGURE 16 A 4,096-pixel resolution image in JPEG 2000 format (line scan digital camera) (20).