Presentation is loading. Please wait.

Presentation is loading. Please wait.

3DVCR Group, Department of Machine Intelligence *Yipu Zhao, M. He, H. Zhao, F. Davoine, and H. Zha Department of EECS, Peking University Sino-French Lab,

Similar presentations


Presentation on theme: "3DVCR Group, Department of Machine Intelligence *Yipu Zhao, M. He, H. Zhao, F. Davoine, and H. Zha Department of EECS, Peking University Sino-French Lab,"— Presentation transcript:

1 3DVCR Group, Department of Machine Intelligence *Yipu Zhao, M. He, H. Zhao, F. Davoine, and H. Zha Department of EECS, Peking University Sino-French Lab, CNRS & LIAMA

2 3DVCR Group, Department of Machine Intelligence Object discovery from mobile laser scanning.

3 3DVCR Group, Department of Machine Intelligence Different applications may concern different objects. Put more focus on the objects of interest.

4 3DVCR Group, Department of Machine Intelligence Objective: Compute the object-based saliency of laser points Computing Object- based Saliency This Research Laser Points Object Detection Geometric Feature Extraction Object Candidate Generation Object-based Saliency Computing Step1Step2Step3

5 3DVCR Group, Department of Machine Intelligence Experimental Platform LMS GPSIMU LMS

6 3DVCR Group, Department of Machine Intelligence Four types of geometric features Vertical lineHorizontal line Vertical planeHorizontal plane Seed Selection Region Growing Range Image Geometric Features Step 1. Geometric Feature Extraction

7 3DVCR Group, Department of Machine Intelligence Extraction results Vertical Line Vertical PlaneHorizontal Plane Horizontal Line

8 3DVCR Group, Department of Machine Intelligence Voting Candidate Centers Clustering Centers Object Candidates Geometric Features

9 3DVCR Group, Department of Machine Intelligence Voting car candidate

10 3DVCR Group, Department of Machine Intelligence The object-based saliency depends on Type & size of the related geometric features Spatial relationship between geometric features To contain these information A graphical object representation is introduced Graph Generation Graph Matching Salient Objects Object Candidates

11 3DVCR Group, Department of Machine Intelligence Graph definition: Node: Type & size of geometric features Edge: Spatial relationship of different geometric features i j x z y Object coordinate i j

12 3DVCR Group, Department of Machine Intelligence Some model graphs of objects of interest CarBusRoad lampTraffic lightTraffic sign

13 3DVCR Group, Department of Machine Intelligence

14 1. Highway scene (the 4 th ring road, Beijing) Collecting time cost: 35 minutes Data volume: about 14,300,000 laser points Sample: 26 model graphs for 8 object classes Processing time: 18 minutes (on a 2.8GHz & 8G PC)

15 3DVCR Group, Department of Machine Intelligence Road lamp Traffic light Traffic sign Road belt Car Signpost

16 3DVCR Group, Department of Machine Intelligence Bus Road lamp Traffic light Traffic sign Road belt Building

17 3DVCR Group, Department of Machine Intelligence 2. Street scene (Street ShangDi, Beijing) Collecting time cost: 30 minutes Data volume: about 13,210,000 laser points Sample: 38 model graphs for 11 object classes Processing time: 20 minutes (on a 2.8GHz & 8G PC)

18 3DVCR Group, Department of Machine Intelligence Truck SignpostCar Building Road lampAd board Result in street scene Data GraphModel Graph

19 3DVCR Group, Department of Machine Intelligence Traffic signBus Building Data GraphModel Graph

20 3DVCR Group, Department of Machine Intelligence Road lamp Traffic sign Road belt Bus Signpost Car Trash boxBuilding

21 3DVCR Group, Department of Machine Intelligence *Highway scene only

22 3DVCR Group, Department of Machine Intelligence Compute the object-based saliency of urban laser sensing data Highlight the data of objects of interest Help object detection in the subsequent procedures In the future On-line application More comprehensive approach (include context information)

23 3DVCR Group, Department of Machine Intelligence Thanks


Download ppt "3DVCR Group, Department of Machine Intelligence *Yipu Zhao, M. He, H. Zhao, F. Davoine, and H. Zha Department of EECS, Peking University Sino-French Lab,"

Similar presentations


Ads by Google