Abstract:
Provided is a map building and position recognizing method using a mobile robot, wherein the method comprises: a providing step of providing the mobile robot including a detecting part, a storing part, a calculating part calculating plane features from point data, a controlling part connected to the calculating part and applying a calculation control signal, and a driving part moving depending on the driving control signal of the controlling part; a detecting step of allowing the detecting part to detect the point data for a peripheral area depending on the detection control signal of the controlling part; a plane feature extracting step of allowing the controlling part to extract the plane features using the point data detected by the detecting part and predetermined data in the storing part; and a data associating step of matching map data for a peripheral environment using a space relation derived from prior plane features and present plane features, wherein the map data for the peripheral area is obtained by the motion of the mobile robot, and the data associating step includes: a sampling step of selecting a pair of rotation matching plane features between the present plane features and the prior plane features; and a consensus step of estimating the space relation between the prior position of the mobile robot and the present position of the mobile robot using the pair of rotation matching plane features and identifying whether or not the pair of rotation matching plane features is included in a group of pairs of the plane features.
Abstract:
The present invention is to provide a plane extracting method, and a map building and position recognizing method having the same and using a mobile robot, wherein the map building and position recognizing method comprises: a providing step of providing the mobile robot including a detecting part, a storing part, a calculating part, a controlling part, and a driving part; a detecting step of allowing the detecting part to detect point data for a peripheral area depending on the detection control signal of the controlling part; a plane feature extracting step of allowing the controlling part to extract plane features using the point data detected by the detecting part and predetermined data in the storing part; and a data associating step of matching map data for a peripheral environment using a space relation derived from prior plane features and present plane features, wherein the map data for the peripheral area is obtained by the motion of the mobile robot. In the plane feature extracting step, a potential cue sequentially storing point data with cube units and data for the corresponding cube units, and a plane cue receiving the data for the cube units from the potential cue are used so that a plane extracting process for hierarchically and differentially partitioning the point data or integrating the plane features depending on the complexity of the peripheral environment is possible, wherein the point data with the cube units is generated by partitioning the point data for the peripheral environment into predetermined cube sizes, and the data for the cube units has single plane features after sequentially processed in the potential cue.
Abstract:
형태론적 연산에 기반한 공간 기하 정보의 노이즈 제거 방법을 제공한다. 본 발명의 일 실시예에 따른 형태론적 연산에 기반한 공간 기하 정보의 노이즈 제거 방법은 대상 영역에 대한 공간 기하 정보에 기초하여, 상기 대상 영역을 소정의 단위로 분할한 복수의 복셀(voxel) 각각에 대응되는 복셀값을 설정하는 단계; 상기 복수의 복셀 각각에 대하여 상기 복셀값과 적어도 하나의 구조적 요소(structuring element)를 이용하는 복수의 형태론적 연산을 수행한 결과로부터, 상기 복수의 복셀 중에서 복수의 제1 노이즈복셀을 선별하는 단계; 및 상기 복수의 제1 노이즈복셀에 기초하여, 상기 공간 기하 정보로부터 노이즈를 제거하는 단계를 포함한다.