Street Sparse Points Reconstruction From Driving Recorder Data And The Applications In Pedestrain Crossing Detection

The driving recorder is a dashboard camera that collects images while the motor vehicle is moving. Utilizing these image data can reduce street scene reconstruction and updating costs because of their low price, wide use, and extensive shooting coverage, the dissertation therefore proposed an automatic region detection based method to reconstruct street scenes from driving recorder images. Then with the help of reconsruction results, pedestrian crossing accurate detection and analysis method is described afterwards.An enormous number of moving vehicles are included in the collected data because the typical recorders are often mounted in the front of moving vehicles and face the forward direction, which can make matching points on vehicles and guardrails unreliable. The dissertation therefore proposed a region detection based method, which use vehicle and guardrail regions mask to improve the structure results from the motion reconstruction. Note that we define vehicle and guardrail regions as "mask" in this paper since the features on them should be masked out to avoid poor matches. After removing the feature points in our new method, the camera poses and sparse 3D points that are reconstructed with the remaining matches. With the structure from motion (SfM) reconstruction methods, the reconstruction results show that the mask decreased the root-mean-square error (RMSE) of the pairwise matching results, which led to more accurate recovering results from the camera-relative poses. The vehicle and guardrail mask also increased the accuracy of point clouds and corrected the problems of the typical methods on repeatedly reconstructing several buildings when there was only one target building.Pedestrian crossing, as an important part of transportation infrastructures, serves to secure pedestrians' lives and possessions and keep traffic flow in order. Since pedestrian crossing is subject to wear and tear from heavy traffic flow, it is of great importance to monitor its location and status quo. On this account, this dissertation puts forward a method based on the reconstruction result which can automatic detect pedestrian crossing and analyse its defilement and impairment. The method firstly trains pedestrian crossing classifier with low recall rate, and then refines those initial detections by utilizing projection filtering and contour information analysis. Finally, a pedestrian crossing detecting method with high recall rate, precision and robustness will be achieved. This method works for pedestrian crossing detection under different situations and light conditions. Based on the features extract from detection results, a pedestrian crossing analysis method is proposed. It can detect and separate defilement\impairment crossings from general ones automatically.