- 102 - PHOTOGRAMMETRIC ADVANCED DETECTION SOLUTION INCIDENT DETECTION IN TUNNELS de Kok P., Zenz T., Markus J. Siemens Building Technologies Austria ABSTRACT Use of photogrammetric methods and image analysis for automatic monitoring and measurement systems. Interesting new detection and evaluation possibilities can be developed on the basis of the combination of photogrammetric measurement methods and image analysis. The following will be explained: 1) the functionality of such measurement methods; 2) possibilities in traffic safety; 3) possibilities in traffic monitoring; and 4) the schematic of such a system. 1. THE FUNCTIONALITY OF SUCH MEASUREMENT METHODS Each single frame of every video stream is handled in the photogrammatic image analysis process. In the first step, the regions of interest are specified. In the second step, the different detection methods of the image analysis processes are used to filter the image areas. In steps three to five, these areas are calculated using photogrammetric algorithms. The employed cameras, computers and algorithms allow these measurements to be completed in real time. Dynamic processes can be calculated for the complete area using a timestamp and the scaled positioning of the cameras. The steps of this process are illustrated in Figure 1. Step 1 Step 3 Step 4 Step 5 Step 2 Figure 1: Process steps
- 103-2. POSSIBILITIES IN TRAFFIC SAFETY Calculation and graphical presentation of speed flow and delay detection through the complete tunnel With the combination of image analysis and photogrammetry, vehicles can be detected, qualified, and their movements exactly calculated. The results can be saved on a time line and graphically displayed for a better understanding. Figure 2 By connecting multiple cameras and their processes, the traffic flow can be calculated, displayed and archived for the complete tunnel area. Every vehicle that drives through the tunnel can be analysed. Figure 3
- 104 - Detection of illegal driving, detection of stopping or slow-driving cars With the help of the movement and direction calculated for all vehicles, illegal driving can be detected directly and the corresponding risk management measures initiated. Figure 4 Detection of dangerous objects in the traffic lanes (lost goods, animals, etc) The video analysis system detects dangerous objects automatically and allows the corresponding risk management measures to be started immediately. Detection of vehicles with dangerous goods Using specially equipped cameras, vehicles with dangerous goods can be detected automatically and their passage through the tunnel monitored as explained above. Classification of vehicles With the help of photogrammetry, the dimensions of the vehicles can be calculated automatically and classified accordingly. Smoke detection The video analysis system detects smoke automatically and allows the corresponding risk management measures to be initiated automatically. Figure 5
- 105-3. POSSIBILITIES IN TRAFFIC MONITORING Calculation of speed and distances, generation of court-admissible files and documentation License plate recognition system to search for listed cars and risk management With the integration of licence plate recognition systems, traffic enforcement cases can be handled fully automatically. Figure 6 Monitoring of prohibited driving and overtaking incidents and generation of courtadmissible files and documentation Identifying illegal drivers With the procedure described above and a special time code method, it is possible to produce court-admissible files and documentation. The drivers are recorded with additional cameras and the license plate recognition is completed automatically. Figure 7
- 106 - SAMPLE SYSTEM SCHEMATIC One example of an analytic system is shown in Figure 8. Figure 8: Siemens SiADSplus system