DIGITAL VIDEO GIS REFERENCED SYSTEM FOR SPATIAL DATA COLLECTION AND CONDITION ASSESSMENT TO ENHANCE TRANSPORTATION ASSET MANAGEMENT

Transcription

1 DIGITAL VIDEO GIS REFERENCED SYSTEM FOR SPATIAL DATA COLLECTION AND CONDITION ASSESSMENT TO ENHANCE TRANSPORTATION ASSET MANAGEMENT Edmundo J. Botner 1* and Mario S. Hoffman 1 1 YONA - Engineering Consulting & Management Ltd., Israel * Corresponding Author s ebotner@yonaltd.com ABSTRACT This paper presents an integrated digital video-gis referenced system for spatial data collection and evaluation. The system, called YonaPMS.Video, uses a new approach that combines high resolution digital video with powerful Geographic Information Systems (GIS) to generate accurate data flows for transportation asset inventory, condition assessment, and for comparing historical video databases essential in modern Transportation Asset Management (TAM) systems. YonaPMS.Video uses a linear referenced GIS model that assigns the road kilometer reference to each video frame in addition to the commonly reported X, Y, Z coordinates. This feature enables the performance of multi-camera virtual trips at the office to produce efficient and accurate hardware spatial surveys of pavements, guardrails, posts, signs, culverts, ditches, bridges, etc suited to the managerial needs of transportation agencies responsible for the maintenance and operation of streets, roads, airports, and railways. In addition, YonaPMS.Video enhances the Quality Assurance/Quality Control (QA/QC) of the collected data as both the geographic position and the element attributes are recorded simultaneously. With YonaPMS.Video it is easy to generate GIS layers directly from the video display which are compatible with most available GIS platforms used in the industry. YonaPMS.Video provides the user with an automatic form generator that enables him to create his own inventory or condition surveying forms suited to his preferences and needs. YonaPMS.Video incorporates a 3-D measuring and positioning tool to make horizontal and vertical measurements and positioning of infrastructure elements directly on the video display. With this 3-D tool it is possible to measure the width of road lanes, the height of guardrails and signs, the area of a patch, and map the coordinates, the road-station kilometer and the attributes of the surveyed hardware elements. YonaPMS.Video has been used in Israel and overseas in numerous projects dealing with pavement maintenance management, concession bidding and operation, and asset inventory and management for national, municipal and privately operated road and airport networks. KEY WORDS: Video-GIS, Transportation Asset Management (TAM) systems, Quality Assurance/Quality Control INTRODUCTION It has long been recognized that a systematic data management is needed to achieve sound, economical and balanced decisions when limited budgets are allocated for the maintenance and rehabilitation of transportation networks. Finn (1) indicates that it is hard to say when the idea started, but probably 50 years ago, some unknown engineer in some unknown state

2 Botner and Hoffman decided he needed to measure the condition of the pavement to better prioritize his maintenance activities, so he made some notes on a piece of paper about the condition of the road, who lived along the route, and what businesses were affected. Since then, the concepts and tools for pavement and asset management made a significant progress. The 1977 book on Pavement Management Systems (PMS) by Haas and Hudson (2) is probably the first publication presenting an orderly treatment of the subject. The wide use of the Pavement Condition Index (PCI) that led to the development of the micro-paver software for highways and airports management developed in the 80's (3) significantly contributed to the systematic treatment of transportation data for decision making, budget allocation, and information management. It was soon recognized that the huge amounts of tabular data collected for pavement management was better displayed and interpreted adding a Geographic Information System (GIS) platform. The NCHRP Synthesis of Highway Practice 335 of 2004 (4), dealing with pavement management applications using GIS, reported that most United States Departments of Transportation (US DOTs) were either using or planning to use GIS or other spatial technologies to support pavement management activities. The synthesis indicated that GIS and other spatial tools facilitate output presentation, data collection and processing, data integration, and incorporation of spatial data into PMS analysis. The concept of pavement management was further broadened. The office of Asset Management, created by the U.S. Federal Highway Administration (FHWA) in 1998, soon published the "Asset Management Primer" (5). It was felt that a systematic process of maintaining, upgrading, and operating physical assets cost-effectively was needed, and asset management provided the tools for doing it. Asset management goes beyond the traditional management practice of examining singular systems within the road network, like pavements or bridges, and looks at the universal system of a network of roads and all of its components to allow comprehensive management of vast infrastructure with limited resources. The NCHRP Synthesis of Highway Practice 371 of 2007 (6), dealing with managing selected transportation assets like signals, lighting, signs, pavement markings, culverts, and sidewalks, recently examined US transportation agencies level of application of the concepts, methods, and tools of asset management. The synthesis indicates that while relatively sophisticated management systems and other analytic tools enable agencies to track the condition and performance of pavements and bridges, these methods and tools are not as widely available or deployed for other surface transportation assets in the US. The synthesis concluded that the state of knowledge regarding the performance and service life of selected assets needs to be improved. Many agencies view the lack of a complete, accurate and current inventory of these selected assets as one of the key issues to address. The incorporation of TAM systems promoted the development of inventory tools that added video images of the surveyed elements, but few papers have been published describing the integration of the video and their corresponding geographic information (7). A big deal of effort has been directed towards the automatic inspection of pavement cracking using video images (8, 9) but this aspect seems to be isolated from the wide asset management concept. Other tools, developed by commercial organizations, have attempted to perform some kind of integration, but the report of the scope and the methods used for doing it are limited. In most cases, the video data are collected using a Global Positioning System (GPS) reference, but there is no full GIS integration to access and display the data. This paper presents the main components of YonaPMS.Video, a computerized system that was developed to integrate tabular, GIS and video synchronized data to facilitate and improve TAM. In particular, YonaPMS.Video helps improve two major components of TAM: a) asset inventory, and b) condition assessment. In addition, YonaPMS.Video exportable products

3 (GIS layers, database tables, etc.) are fully compatible with the systems commonly used by transportation agencies. YonaPMS.Video database is composed by digital video images obtained from a moving vehicle equipped with multiple cameras, integrated with a GIS referenced map. From the video images it is possible to measure within a ± 4% accuracy, horizontal and vertical elements in the front and rear views. The positioning of infrastructure elements directly from the video can be done accurately depending on the GPS and the perspective model precisions. It is also possible to conveniently perform inventories and condition surveys using custom made forms created by the user with the YonaPMS.Video automatic form generator. These features make YonaPMS.Video an optimal tool for efficient network level inventories and condition surveys. One of the major innovations of YonaPMS.Video is the full GIS controlled access to the database with a simple click on the selected point on the map. With YonaPMS.Video there is no need to access the road information using a tabular data entry form. With the GIS control of the video database, YonaPMS.Video enhances the efficiency and productivity rates of the surveyors, and facilitates access to the data by managers and decision makers. In addition, since all data are geographically and temporally referenced, YonaPMS.Video allows viewing synchronized video images of the same road sections taken at different times for historic comparison. FIELD EQUIPMENT YonaPMS.Video field data collection can be configured using a flexible array of cameras ranging from one simple digital camera mounted on a non-dedicated vehicle, to a multicamera setup covering up to 360 º. The survey vehicle is equipped with the following typical components: 1. A high resolution digital video system (color and/or B/W) comprising front panoramic and rear cameras perpendicular to the road surface. As noted, the quantity and configuration of the camera setup can be changed to suit specific project needs. 2. A GPS receiver connected to the main PC transmitting according to the NMEA-0183 (National Marine Electronic Association) communication protocol. The GPS receiver registers the survey vehicle location as a function of time. GPS precision is generally a function of the project requirements. Current sub-meter accuracy GPS receivers cost less than US$ 3,000 while lower precision GPS are readily available at very low cost and can achieve precisions of about 15 meters that may be sufficient for some project needs. 3. A Distance Measurement Instrument (DMI) to accurately record the trip distance. 4. A main PC in charge of: a) sending a trigger to the cameras to get a simultaneous set of images as a function of the distance traveled (every 5, 10 meters, etc.) selected by the user, b) Create a file called GPS log containing the GPS information (latitude, longitude, altitude, UTC Time, speed and bearing), and c) Create a file called DMI log, registering the trigger time and the trip distance. 5. A Cameras PC; one or more computers (depending on the project requirements) in charge of: a) controlling the cameras settings (shutter speed, gamma, brightness, etc.), b) record the video images, c) create a file called Cameras log that assigns to each video frame an ID name and a recording time. 6. The Software YonaPMS.RVS that controls and monitors the system functioning.

4 Botner and Hoffman All vehicle systems are synchronized with respect to time. For convenience, the UTC time provided by the GPS receiver is selected as the system time (ST). DATA INTEGRATION PROCESS Figure 1 shows the data integration process which connects and relates the field data with the YonaPMS.Video environment. FIELD DATA YonaPMS.Video GPS logs DMI logs Cameras logs RAW GPS (X, Y, Z) = f (ST) RAW Image Image = f (ST) Image Database Image = f (camera,x, Y, Z, Road, km, ST) Video Records Video Storage GIS FIGURE 1 Field Data Integration Process Raw GPS data file Using the GPS logs file, a RAW GPS file is created that contains the latitude, longitude, altitude, bearing, speed, and UTC time. This file incorporates projected X, Y, Z coordinates according to the local coordinate system. Raw Image data file There is a short time gap between the moment the trigger is activated and the moment the video frames are saved in the hard disk. This time gap is used by the camera sensors to expose and integrate the images, send them to the PC memory RAM, and save them to the hard disk. In addition, the storing time of each image taken simultaneously is different since the saving time to the hard disk is sequential, i.e. one picture at the time. To overcome this problem and assign to each frame its true exposed time, a compatibility process is implemented between the data in the DMI log and the Cameras log files to generate a new synchronized file called RAW Image Data. Connecting data files

5 The Raw GPS data file and the Raw Image data files are connected using linear interpolation based on the system time. As a result of this interpolation, each video frame gets its true position and time coordinates: X, Y, Z and the ST. Assignment of road kilometer reference The last step consists in assigning a road kilometer reference to each video frame. This is done using a linear referencing technique (10). For this purpose, a GIS layer containing a linear reference of the road kilometer (milestones) is needed. Thus, each vertex of the road must contain the kilometer reference in addition to its X-Y coordinates. Mapping each frame according to its coordinates, and projecting this point onto the road axis shown in a GIS map, a kilometer reference is determined. At this point, a full description of each video frame is obtained containing the information described in the following relation: Video Image i = F (camera, X, Y, Z, Road, km, ST) [1] 3D-MEASURING AND POSITIONING TOOL FUNDAMENTALS Figure 2 shows the YonaPMS.Video 3-D measuring tool scheme. It is based on a geometric model representing the perspective of an observer located at the camera lens focus. OBSERVER Interpretation Plane FIGURE 2 Observer perspective scheme Model Parameters Table 1 describes the model parameters and their use in the 3-D model.

6 Botner and Hoffman TABLE 1 Model Parameters No. Parameter Symbol Unit Use 1 Lens horizontal angle of view β deg 2 Lens vertical angle of view δ deg Measurement & Positioning Measurement & Positioning 3 Vehicle trip direction relative to the North θ deg Positioning 4 Horizontal distance between the GPS antenna and the cameras axis D m Positioning 5 Camera lens height relative to the Interpretation Plane H m Measurement & Positioning 6 Camera lens axis angle relative to the Interpretation Plane α deg Measurement & Positioning Figures 3 and 4 illustrate the model parameters. OBSERVER LATERAL VIEW H(m) δ( ) α( ) INTERPRETATION PLANE OBSERVER β( ) INTERPRETATION PLANE TOP VIEW FIGURE 3 Lateral and Top view of the model parameters

7 FIGURE 4 Vehicle model parameters Determination of Model Parameters 1. Parameters 1 and 2 in Table 1 are obtained directly from the technical specifications of the lens producer. 2. Parameter 3 is obtained from the GPS bearing representing the angle of the trip direction relative to the North which is defined as 0º. 3. Parameter 4 represents the measured horizontal distance between the GPS receiver and the cameras axis. 4. Parameters 5 and 6 are determined using a calibration process as described below. Calibration to determine H and α Once a week or when the cameras have been re-installed, a calibration is made to determine the parameters H and α. To this effect, about 10 wooden sticks of 1 meter length are randomly placed on a smooth planar floor near the stopped vehicle and filmed. Figure 5 shows a typical setup of the wooden sticks on the floor and the calibration form. The first step of the calibration consists in assigning an approximate initial value of H and α called H 0 and α 0. Then, the sticks are measured with the YonaPMS.Video 3-D measuring tool starting with the initial H 0 and α 0 values. These initial values are corrected in successive loops until a predetermined accepted error tolerance is achieved. Normally, a 3% error is accepted.

8 Botner and Hoffman FIGURE 5: Calibration process Examples of 3-D measurements Figure 6 shows examples of horizontal and vertical measurements with the YonaPMS.Video 3-D tool. The vertical accuracy achieved when the measurements are taken close to the viewer in the inferior 2/3 part of the picture reaches levels similar to the horizontal measurements. The error ranges in the order of ±3%. FIGURE 6: Horizontal and Vertical Measuring Examples

9 SPATIAL DATA COLLECTION PROCESS The spatial data collection process begins with a "virtual trip" at the office using YonaPMS.Video. Figure 7 displays the main screen which is composed of two windows: a) the GIS window that shows the network map linearly referenced as previously described, and b) the VIDEO window that shows the video images synchronized with the vehicle position shown in the GIS map. On the GIS window it is possible to display additional GIS layers either in vectorial format (shp files or dwg files) or raster files (orthophotos, etc.).when the video data capturing is made with a multi-camera configuration, the VIDEO window can display each camera separately, or in a compounded panoramic view as shown in Figure 7. VIDEO Window VEHICLE LOCATION GIS Window FIGURE 7 YonaPMS.Video main screen Since YonaPMS.Video knows the spatial and temporal position of the survey vehicle for each one of the video images, and the 3D-measuring tool can determine the local coordinates X P, Y P, Z P, Road P, KM P, ST of each of the video pixels in the image, YonaPMS.Video can build a new GIS layer containing road elements appearing on the VIDEO window using the survey forms generator explained below. Survey Forms Generator YonaPMS.Video contains an automatic survey forms generator (SFG) that gives the user ample freedom to build his own surveying forms according to his needs and preferences. Once generated, the form is saved by YonaPMS.Video and used by the surveying team for

10 Botner and Hoffman entering data and performing the survey. Changes in the forms can be made by the administrator at any time during the data survey as new conditions or requirements arise. A typical survey form generated by the SFG includes the following basic components: 1. Survey Name: A name describing the survey. 2. GIS Element Type: Line or point elements. 3. Element Position: Element position relative to the lateral cross section of the road, i.e. if the surveyed element is the guardrail, for instance, the possible element position can be right (R), left (L), Center (C), etc. 4. Element Attributes: There are two types of attributes: a) System attributes, and b) User defined attributes. System attributes cannot be changed as they include the location and the time of the video frame furnished by the system. User defined attributes are defined by the user among the following types: a. Combo box: a list to choose from predetermined options. b. Text/Numeral: text or numerals entered by the user as needed. c. Check box: True/False when there are just two options. d. Picture: Option to save pictures from the VIDEO window. 5. Combo box options: Values entered into the combo box. Form Generator Example The following example illustrates the use of the SFG to generate a form to collect road signs data. 1. Survey Name: Survey of Road Signs 2. GIS Element Type: Point. 3. Element Position: Left (L), Center (C) and Right (R). 4. Element Attributes: see Table 2 TABLE 2 Element Attributes Types and Values for a Road Signs Survey Element Attribute Name Type Value/ Combo box options The Sign is permanent Check box T/F Sign Type Combo box Regular Simple Frame Electronic Frame Simple Cantilever Electronic Cantilever

11 Element Attribute Name Type Value/ Combo box options Sign Subtype Belongs To Height (m) Width (m) Picture Combo box Combo box Text/Numeral Text/Numeral Picture City Entrance Driver Information General Information Road Number Municipality PWD Private Concession Operator Other Measured value from the VIDEO window Measured value from the VIDEO window Attach picture from the VIDEO window Data collection and Interpretation Data collection and interpretation is done by trained surveyors using a computer station with preferably two monitors, one showing the GIS window and the data form, and the other showing the VIDEO window. For example, the sequence of data collection and interpretation regarding the signs example in Table 2 is as follows: 1. Select and open as many forms as "element positions" are possible. In this case of road signs the positions are 3 for L, C, and R. 2. With a "click" on the GIS window, position the beginning of the survey at the selected road and kilometer and start the "virtual trip" at a selected speed. 3. Stop near the first sign appearing on the VIDEO window, and choose the best camera view of the sign. 4. Click on the sign base to automatically register the System Attributes (X, Y, Z, Road, KM, ST) of the sign. 5. Fill in the form fields and click "apply" to save the data on the database once finished. At this step, a point representing the sign position is automatically added into the "road signs" GIS layer. To edit and view the sign information and attributes at any time, just click on the point in the map. 6. Continue the "virtual trip" until the next sign appears, and repeat the process. Figure 8 shows a YonaPMS.Video screenshot of the sign data survey entry form.

12 Botner and Hoffman FIGURE 8 YonaPMS.Video screenshot of the sign data survey SUMMARY AND CONCLUSIONS This paper presented the YonaPMS.Video system for spatial data collection and analysis to enhance Transportation Asset Management (TAM) activities. The system uses a new approach that combines high resolution digital video with powerful Geographic Information Systems (GIS) to generate accurate data flows for asset inventory, condition assessment, and for comparing historical video databases essential in modern TAM. YonaPMS.Video can be operated using a flexible array of cameras ranging from one simple digital camera mounted on a non-dedicated vehicle, to a multi-camera setup covering up to a 360 º panoramic view and vertical cameras to view the pavement surface for condition evaluation. The full capabilities of the system can also be obtained in limited-budget projects using just one camera. YonaPMS.Video incorporates a linear referenced road network model into a fully compatible GIS environment facilitating the execution of inventory surveys and asset condition evaluation of guardrails, posts, signs, culverts, ditches, bridges, etc increasing the data collection efficiency and reliability. With a simple click on the GIS map the surveyor can easily access any section of the road network to display video and database information, improving the tabular accesses commonly used by other systems. YonaPMS.Video incorporates a 3-D measuring and positioning tool to make horizontal and vertical measurement and positioning of infrastructure elements directly from the video pictures. With this tool it is possible to measure the width of road lanes, the height of guardrails and signs, the area of patching zones, and map the coordinates, the road-station kilometer and the attributes of the surveyed hardware elements. YonaPMS.Video provides the

13 user with an automatic form generator that enables him to create his own inventory and surveying forms suited to his preferences and needs. YonaPMS.Video has been used in Israel and overseas in numerous projects dealing with pavement maintenance management, concession bidding and operation, and asset inventory and management for national, municipal and privately operated road and airport networks. The tangible benefits obtained from a fully integrated tabular-video-gis database generated with YonaPMS.Video far overshadow the low costs associated with building and maintaining these databases. REFERENCES 1. Finn, F., "Pavement Management Systems Past, Present, and Future", Public Roads, Vol. 62, No. 1, July/August Haas, R. C. G., and Hudson, W. R., "Pavement Management Systems", McGraw-Hill, New York, NY, Shahin, M. Y., and Kohn, S. D., "Pavement Maintenance Management for Roads and Parking Lots", Technical Report M-294, Construction Engineering Research Laboratory, U.S.A.C.E., October "Pavement Management Applications Using Geographic Information System", NCHRP Synthesis 335, "Asset Management Primer". FHWA, U.S. Department of Transportation, "Managing selected transportation Assets: Signals, Lighting, Signs, Pavement Markings, Culverts, and Sidewalks", NCHRP Synthesis 371, Santiago-Chaparro, K. R., Colucci-Rios, B., and Figueroa-Medina, A., M., "Development and Evaluation of a Software Tool that Integrates GPS and Video Data from a Road Alignment to Perform Road Condition, Safety Audits, and Inventory Surveys", Presented at the 87 th Annual Meeting of the TRB, Washington DC, Huang, Y., and Bugao, Xu, "Automatic Inspection of Pavement Cracking Distress", Report No. FHWA/TX-06/ , Center for Transportation Research, Texas DOT, Offrell, P., Sjogren, L., and Magnusson, R., "Repeatability in Crack Data Collection on Flexible Pavements: Comparison Between Surveys Using Video Cameras, Laser Cameras and a Simplified Manual Survey", Journal of Transportation Engineering, Vol. 131, No. 7, American Society of Civil Engineering, Linear Referencing and Dynamic Segmentation in ArcGIS 8.1. An ESRI white paper, MetaID=294, Accesed February 20 th, 2011.