{"created":"2023-05-15T11:55:19.043190+00:00","id":190,"links":{},"metadata":{"_buckets":{"deposit":"cc9254e0-c50a-4798-a8d0-2efbc8a58090"},"_deposit":{"created_by":18,"id":"190","owners":[18],"pid":{"revision_id":0,"type":"depid","value":"190"},"status":"published"},"_oai":{"id":"oai:kyutech.repo.nii.ac.jp:00000190","sets":["6:7"]},"author_link":[],"item_20_date_granted_61":{"attribute_name":"学位授与年月日","attribute_value_mlt":[{"subitem_dategranted":"2007-03-23"}]},"item_20_degree_grantor_59":{"attribute_name":"学位授与機関","attribute_value_mlt":[{"subitem_degreegrantor":[{"subitem_degreegrantor_name":"九州工業大学"}],"subitem_degreegrantor_identifier":[{"subitem_degreegrantor_identifier_name":"17104","subitem_degreegrantor_identifier_scheme":"kakenhi"}]}]},"item_20_degree_name_58":{"attribute_name":"学位名","attribute_value_mlt":[{"subitem_degreename":"博士（工学）"}]},"item_20_description_30":{"attribute_name":"目次","attribute_value_mlt":[{"subitem_description":"1 Introduction||2 Overview of sewer systems and inspection systems||3 A vision-based automated fault detection system||4 Navigation based on single camera and IR sensors||5 Navigation based on stereo camera and laser scanner||6 Conclusions||Bibliography","subitem_description_type":"Other"}]},"item_20_description_4":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"(Abstract)Pipe walls in sewer systems are prone to be damaged due to aging, trafficand chemical reactions, through which inflow such as rainwater and groundwaterseeps into pipe systems. Regional city government reports state thatthis inflow amounts to approximately 30% of the total flow. In additionto the inflow of groundwater into the sewer pipes, outflow from damagedsystems also occurs, contaminating the surrounding environment.Conventional inspection of a sewer pipe system is carried out using acable-tethered robot with an onboard video camera system. This robot isconnected to the outside of sewer system by a cable. The cable is usedfor energy supply, transmission of commands from a human operator to thedevice, data transmission back to an operator, a life-line in case the devicegets stuck in the pipe, and measuring the distance traveled. An operatorremotely controls the movement of the robot and the video system. By thisvideo-supported visual inspection, any notable damages or abnormalities arerecorded in video stream. The reliability of this system depends on theexperience of an operator. The system is also prone to human error, and tendsto be time consuming and expensive. Consequently, effective autonomousrobot capable of online identification and extraction of objects of interestsuch as cracks from sensory signals is of immediate concern.Based on the above, we design a prototype autonomous mobile robot,KANTARO, for inspecting sewer pipes. It is able to move autonomously in200-300mm diameter sewer pipes, to smoothly turn 90 degrees at a junction,and to go down a step. KANTARO carries all required resources such asa control unit, a camera, a 2D laser and an IR sensor. Damages or abnormalitiesin sewer pipes are detected based on recorded sensory data. In thisthesis, I focus on an automated fault detection system, navigation system,and stereo vision system for autonomous inspection robots such as KANTARO.Robust detection of cracks and other faults in sewer pipes based on sensorydata is another important challenge. However, all related previous worksfocused on specific types of faults in pipes, hence were unable to detect multipletypes of faults. Accordingly, a truly automated fault detection systemis currently not available in the real world. I propose a method for detectingfaulty areas based on images, and an automated intelligent system designedto facilitate diagnosis of faulty areas in a sewer pipes system. The systemutilizes image analysis and efficient techniques for providing the location andthe number of faults in a sewer pipe system. In contrast to the conventional manual system, the proposed system can automatically detect faults andmove in real time. Its detection performance is 100%, when the false positiverate is 34%. This ratio is said to be acceptable for sewer inspection, and thereduction of time and cost is also realized.Another central issue in developing an autonomous sewer robot is itsnavigation. Navigation of an autonomous sewer robot based on a map ofsewer pipe system is not applicable as it is, because large slips in sewerpipes tend to produce erroneous odometry information, causing unreliablelocalization. It is to be noted that data from Global Positioning System(GPS) are not available in underground sewer pipe systems. Accordingly,an autonomous robot has to estimate the current position based on localfeatures.Navigation of an autonomous sewer robot is composed of the followingtasks. Firstly, estimation of the current position based on salient local featuressuch as manholes, inlets and pipe joints. Secondly, finding a path.Thirdly, following the path in the real sewer pipe system. Resulting mapsof the sewer pipe system describe pipes, manholes and other local features,which contribute to localization. I propose a method for navigation of anautonomous inspection robot based on fusion of single camera images andIR sensor data. It is capable of self localization, which cannot be done by theconventional methods. We also conduct experiments for sewer robot navigationin a dry sewer test field at FAIS-RDSO, Kitakyushu. They succeedin detecting local features and show high performance of self localizationby using sensory information. In additional to using a single camera in theabove proposed methods, I also use a stereo camera to see the performanceof stereo vision for navigation, which is described as follows.Stereo matching is an essential issue in computer vision. Recently, manystereo matching algorithms based on segmentation, graph cuts and so onhave been proposed. Because the disparities change continuously in sewerenvironment, these methods are not applicable to sewer systems and arecomputationally expensive. I propose a cooperative stereo matching algorithmusing Sum of Squared Differences (SSD) and Linear Computation (LC)measures, which can be implemented in a real-time system. It is a robustalgorithm for sewer inspection in robot vision. The algorithm produces aneasy-to-understand distance map of the sewer, emphasizing the feature region.The computational time by this algorithm is about 1/5 compared withthat by other algorithms such as the conventional SSD. In order to reducethe computational time, I also propose a fast stereo matching algorithm usinginterpolation. The computational time by the proposed algorithm is only1/20 of those by the conventional algorithms such as the SSD. Hence it issuitable for our real-time sewer vision system.The above stereo matching algorithm is utilized for proposing anothermethod for navigation which is based on stereo camera images and laserscanner data. Experimental results of self localization show high performancein providing the appropriate distance. We also design a new mobilelaser scanner for KANTARO. The locations of landmarks in sewer pipe systemare estimated successfully based on measurements. The laser scanneris fast enough to continuously scan relevant pipe sections in the presence oflandmarks, while the KANTARO moves at ordinary inspection speed of lessthan 15cm/s. Also moving the KANTARO in our sewer test field by usingthe proposed method is done successfully.","subitem_description_type":"Abstract"}]},"item_20_description_5":{"attribute_name":"備考","attribute_value_mlt":[{"subitem_description":"九州工業大学博士学位論文 学位記番号:生工博甲第58号 学位授与年月日:平成19年3月23日","subitem_description_type":"Other"}]},"item_20_description_60":{"attribute_name":"学位授与年度","attribute_value_mlt":[{"subitem_description":"平成18年度","subitem_description_type":"Other"}]},"item_20_dissertation_number_62":{"attribute_name":"学位授与番号","attribute_value_mlt":[{"subitem_dissertationnumber":"甲第58号"}]},"item_20_identifier_registration":{"attribute_name":"ID登録","attribute_value_mlt":[{"subitem_identifier_reg_text":"10.18997/00000184","subitem_identifier_reg_type":"JaLC"}]},"item_20_text_34":{"attribute_name":"アドバイザー","attribute_value_mlt":[{"subitem_text_value":"石川, 眞澄"}]},"item_20_version_type_63":{"attribute_name":"出版タイプ","attribute_value_mlt":[{"subitem_version_resource":"http://purl.org/coar/version/c_970fb48d4fbd8a85","subitem_version_type":"VoR"}]},"item_access_right":{"attribute_name":"アクセス権","attribute_value_mlt":[{"subitem_access_right":"open access","subitem_access_right_uri":"http://purl.org/coar/access_right/c_abf2"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Alireza, Ahrary","creatorNameLang":"en"}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_date","date":[{"dateType":"Available","dateValue":"2007-11-27"}],"displaytype":"detail","filename":"sei_k_58.pdf","filesize":[{"value":"32.0 MB"}],"format":"application/pdf","licensetype":"license_note","mimetype":"application/pdf","url":{"label":"sei_k_58.pdf","objectType":"fulltext","url":"https://kyutech.repo.nii.ac.jp/record/190/files/sei_k_58.pdf"},"version_id":"c8fb8436-c455-4270-80e4-61b5ab304c26"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"Vision System","subitem_subject_scheme":"Other"},{"subitem_subject":"Autonomous","subitem_subject_scheme":"Other"},{"subitem_subject":"Inspection Robots","subitem_subject_scheme":"Other"},{"subitem_subject":"Fault Detection","subitem_subject_scheme":"Other"},{"subitem_subject":"Navigation","subitem_subject_scheme":"Other"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourcetype":"doctoral thesis","resourceuri":"http://purl.org/coar/resource_type/c_db06"}]},"item_title":"Research on a Vision System for Autonomous Inspection Robots","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Research on a Vision System for Autonomous Inspection Robots","subitem_title_language":"en"},{"subitem_title":"自律型検査ロボットのビジョンシステムに関する研究","subitem_title_language":"ja"}]},"item_type_id":"20","owner":"18","path":["7"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2007-11-27"},"publish_date":"2007-11-27","publish_status":"0","recid":"190","relation_version_is_last":true,"title":["Research on a Vision System for Autonomous Inspection Robots"],"weko_creator_id":"18","weko_shared_id":-1},"updated":"2024-01-11T02:58:49.136588+00:00"}