MINISTRY OF EDUCATION AND TRAINING HANOT UNIVERSITY OF SCIENCE AND TECHNOLOGY NGUYEN VAN GIAP VISION — BASED LOCALIZATION 1D: 2014BMTCT-KH03 MASTER THESIS OF SCIENCE COMPUTER SCIENCE SUPER VISOR: Vu Hai, Ph. Hanor 2016 CỘNG HÒA XÃ HỘI CHỦ NGHĨA VIỆT NAM Độc lập — Tự do — Hạnh phúc BẢN XÁC NHẬN CHỈNH SỬA LUẬN VĂN THẠC SĨ 'Hụ và tên tác giá luận văn : Nguyễn Văn Giáp Để lài luận văn: Định vị sử dụng thông tin hình ảnh Chuyên ngành: Khoa học máy tính Mã số 5V: CB140975 Tae gia, Người hướng dẫn khoa học và Hội đồng chấm luận văn nhiên tác giả đã sửa chữa, bổ sung luận văn theo biên bản họp Liội đồng ngày 21 thang 10 năm 2016 với các nội dung sau: Bé sung trịch dẫn chơ các hình vẽ, bỗ sung thứ nguyên — Sửa các lỗi soạn thảo trang toàn bộ luận văn. — Bê sung thêm đanh mục các ký hiệu, ý nghĩa và thứ nguyên Bé sung cáo thông tin còn thiểu trong tài liệu tham kháo — Lâm rõ việc sit dung lai code, chỉ rô phan nào, ở đầu 1à Nội, ngày. thắng 11 năm 2016 Người hướng dẫn Tác giả liận văn TS.
Va Hai Nguyễn Văn Giáp CHỦ TỊCH HỘI ĐỒNG GS. Eric Castelli DECLARATION OF AUTHORSHIP J, Nguyen Van Giap, declare that this thesis titled, “Vision based localization” and the work presented in it are my own. 1 confirm that, «This work was done wholly or mainly while in carulidature for a research degree at this University « Where any part of this thesis has previously been submitted for a degree or any other qualification at this University or any other institution, this has been clearly stated. ¢ Where I have consulted the published work of others, this is always clearly allributed.
¢ Where I have quoted from the work of others, the source is always given. With the exception of such quotations, this thesis is entirely my own work, « [have acknowledged all main sources of help, @ Where the thesis is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself. Signed: Date: bà Figure 3.16: Position and heiglit of objeot.17: Ground truth datasel. lo train GP model Tigure 3.18: Ileight detection abject (II det) Figure 3.19: Estimated height of object (H_es0.20: Tracking results hìgh objecL with ŒP model seanario† Figure 3.21: Height of object tracking results scenario 2.22: Outliors of tracking object and removal outlicrs result.23: Consensus between H_det and H_est 1 igure 3.24: Result of Kalman filter wsing.
Trackáng results with and wi[loul processing.1: Testing environment Figure 4. HH Hee Figure 4.3: Some (rames of scenario 1 Figure 4.4: Some frames of scenario 2.5; Low quality traolking resglis.6: Mapping moving object result wilh BGS and shadow ramoval Figure 4.7, Compare tracking BGS IL det with IT gt Figure 4. Compare tracking BG and shadow removal H_det with H_gt.10: Results with BGS-Shadow removal and GP Figure 4.11: Results with applicationt > Figure 4.12; Calculation m scenari Figure 4.13: Result with ? scenario 2 Tigure 4.14: Result with F scenario Ì.15: Values of ? scenario Ì. ACKNOWLEDGEMENTS 1am so honor to be here the second time, in one of the finest university in Vietnam to write those grateful words ta people who have been supporting, guiding me from the very first moment when | was an undergraduate stndent until now, whenI am writing my master thesis I am grateful to my supervisor, Dr.Vu Ilai, whose expertise, understanding, generous guidance and suppart made the research topics to be possible for me that were of great interest to me.
I am pleasure ta work with him. I would like to special thanks to Dr. Le Thi Lan, Dr. Tran Thi Thanh Hai and all of the members in the Computer Vision Department, MICA Institute for their sharp comments, guidance for my works which helps me a lot in how to study and to do researching in right way and also the valuable advices and encouragements that they gave to me during my thesis.
Particularly, 1 would Lo express my appreciations toa Ph. Student Pham Thi Thanh Thuy, who allow me to use a valuable dalabase of human tracking im a surveillance eamera network. Without her permission, T could nol make exlensive evaluations for the proposed method Finally, T would especially like lo thank my family and friends for their continue love, support they have given me through my life, helps me pass through all the frustrating, struggling, confusing. Thanks for everything that helped me get to this day.
Hanoi, 10/2016 Nguyen Van Giap w positioning in indoor environments. ‘these evaluations are implemented in several indoor enviroments. HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY International Research Institute MICA Computer Vision Department Master of Science Vision — Based Localization by Nguyen Van Giap Abstract Nowadays, the vision-based localization systems using vision sensors are widely used in public and crowded places. Positioning information, which is extracted from the image data stream by surveillance cameras, could support monitoring services in different manners.
For instance, to detect people/subjects who are not allowed entrancing in a certain place; or to link human trajectories and then to identify their abnormal behaviors, so on. These services always require positioning information of the interested subjects. Whereas, the other localization techniques are still limited about distance, accuracy (e., WIFI, RFID), or setting the environment and usability (e. The vision-based localization technique, particularly, in indoor-environment has many advantages such as: scalable, high accuracy; without requirements of the additional attached- equipment/devices to the subjects.
Motivated by above advantages, this thesis aims to study, and propose a high accuracy vision-based localization system in indoor environments. We also take into account detailed implementations and developments, as well as testing the proposed techniques. It is noticed that the thesis focuses on moving human indoor environments that is monitored in a surveillance network camera. To archive a high accuracy positioning system, the thesis deals with the critical issues of a vision-based localization.
We observe that there is no a perfect human detector and tracking. Then, we use a regression- model to eliminate outlier detections. The system therefore improves detection and tracking results. Throughout the thesis, we first briefly introduce an overview of vision — based localization, We then present the proposed frame-work including steps: Background Subtraction for detecting moving subject; shadows removal techniques for improving detection result, and linear regression method to eliminate the outliers; and finally the tracking object using a Kalman Filter.
The most important result of the thesis is demonstrations which show a high accuracy and real-time computation for human 3 LIST OF FIGURES Figure 1.1: Tadoor localization techniques.2: Surveillance camera network.3: Positioning a human fiom video stream.4: Casting shadow problem. a) Origin image: (b)-(c) Casting shadowa, đ) Mask of abject; (e) Shadow pixels - - 14 igure 1.5: Some examples of the experimental environmenis. a) hallway, b) Lebby and c) in a room. ‘The flow chart ofa common vision-based localization technique.
Kinds of object shadows ([9]) 23 Figure 2.3: An illustratrion of the human tracking results in [9].4: Different tracking approaches.1: Foot-point definition.2: ‘fransformation 2D point to 3-D point real world. Top row: The original image; Bollom row: the corner poinls are detected for ealculalinys the homographic matrix. A wrong tracked-point detection .5: The general flow char† of the proposed method.6: Reaull of BGS with adaptive Gaunsian Mixture. shadow: a) Ongin situation; b) Mask situation 30 Figure 3.8: An illustration of the wrong object detection results due to shadow ApPĐoaraitces.9: Noisy by illuninaHion ehmnging .11: Example using chưomatic-based features for shadow remova.13: Problems of shadow removal.14; Graphical model of the Gaussian Rogression.15: Position and height of object.
38 8 positioning in indoor environments. ‘these evaluations are implemented in several indoor enviroments. LIST OF FIGURES. LIST OF TABLES.
Context and Motivation. Vision-hased localization and main contributions.4, Scope and limitations on the research. Chapter 2: RELATED WORK. A briefly survey on localization techniques.
A brief servey on vision-based localization systems. Chapter 3: PROPOSED FRAMEWORK. Formulate the vision-based laralizatian. Post-processing pracedur 3.4, Shadow removal techniques.
Chromaticity-based feature extraction - - 32 3. Shadow-matching score utilizing physical properties. Existing issues after applying shadow removal. veto neenenen nee eS 3.
A localization estimation using a regression. Definition of Gaussian processing. Estinating hoïglt with regression mođbÌ. Chapter 4: EXPERIMENTAL EVALUATIONS.
4,2, Evaluation results of the BGS and shadow removal. Evaluation results of the Gaussian pracessing regres: 4.1, Evaluation results with GP o. cess ieeeeess ieee eens "—- 4. Evaluation about suitable method 51 4.4, ‘The final cvaluation results of the proposed system.
Chapter 5: CONCLUSION AND FUTURE WORKS. HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY International Research Institute MICA Computer Vision Department Master of Science Vision — Based Localization by Nguyen Van Giap Abstract Nowadays, the vision-based localization systems using vision sensors are widely used in public and crowded places. Positioning information, which is extracted from the image data stream by surveillance cameras, could support monitoring services in different manners. For instance, to detect people/subjects who are not allowed entrancing in a certain place; or to link human trajectories and then to identify their abnormal behaviors, so on.
These services always require positioning information of the interested subjects. Whereas, the other localization techniques are still limited about distance, accuracy (e., WIFI, RFID), or setting the environment and usability (e. The vision-based localization technique, particularly, in indoor-environment has many advantages such as: scalable, high accuracy; without requirements of the additional attached- equipment/devices to the subjects. Motivated by above advantages, this thesis aims to study, and propose a high accuracy vision-based localization system in indoor environments.
We also take into account detailed implementations and developments, as well as testing the proposed techniques. It is noticed that the thesis focuses on moving human indoor environments that is monitored in a surveillance network camera. To archive a high accuracy positioning system, the thesis deals with the critical issues of a vision-based localization. We observe that there is no a perfect human detector and tracking.
Then, we use a regression- model to eliminate outlier detections. The system therefore improves detection and tracking results. Throughout the thesis, we first briefly introduce an overview of vision — based localization, We then present the proposed frame-work including steps: Background Subtraction for detecting moving subject; shadows removal techniques for improving detection result, and linear regression method to eliminate the outliers; and finally the tracking object using a Kalman Filter. The most important result of the thesis is demonstrations which show a high accuracy and real-time computation for human 3 ACKNOWLEDGEMENTS 1am so honor to be here the second time, in one of the finest university in Vietnam to write those grateful words ta people who have been supporting, guiding me from the very first moment when | was an undergraduate stndent until now, whenI am writing my master thesis I am grateful to my supervisor, Dr.Vu Ilai, whose expertise, understanding, generous guidance and suppart made the research topics to be possible for me that were of great interest to me.
I am pleasure ta work with him. I would like to special thanks to Dr. Le Thi Lan, Dr. Tran Thi Thanh Hai and all of the members in the Computer Vision Department, MICA Institute for their sharp comments, guidance for my works which helps me a lot in how to study and to do researching in right way and also the valuable advices and encouragements that they gave to me during my thesis.
Particularly, 1 would Lo express my appreciations toa Ph. Student Pham Thi Thanh Thuy, who allow me to use a valuable dalabase of human tracking im a surveillance eamera network.
