Abstract:
Disclosed is an apparatus for synchronizing stereoscopic images. The apparatus comprises a similarity calculation unit for calculating a similarity between a left image frame and a right image frame or similarity between blocks based on any one frame or frame block of a left image of a stereoscopic image or a right image of the stereoscopic image; a frame difference extraction unit for extracting a frame difference between the left image and the right image based on the calculated similarity between the frames; and a frame synchronization correction unit for synchronizing the left image and the right image by correcting the extracted frame difference.
Abstract:
Disclosed is a method for detecting the existence of a part by analyzing an image obtained by photographing a part of a chassis. The method comprises a step of extracting a first image sample in which a part exists, a second image sample in which the part does not exist, and a third image sample which is irrelevant to the existence of the part; a step of learning how to detect the part from the image using the extracted first, second, and third samples; and a step of determining the existence of the part learned in a predefined interest section of the image, obtained by photographing a part of the chassis, based on the learning results.
Abstract:
PURPOSE: A method for estimating vehicle-to-vehicle distance is provided to minimize error of estimated distance by analyzing a rear image of a preceding vehicle. CONSTITUTION: A method for estimating vehicle-to-vehicle distance comprises: a step of calculating a first estimated distance of a preceding vehicle from an rear image of the preceding vehicle; a step of calculating a scale factor of the preceding vehicle using the first estimated distance and the width of the preceding vehicle; a step of calculating the scale factor in a predetermined time interval and determining the average scale factor of the preceding vehicle; and a step of calculating a second estimated distance using the average scale factor and the width of the preceding vehicle. An apparatus(100) for estimating vehicle-to-vehicle distance comprises a first estimated distance calculating unit(120), a scale factor calculating unit(130), an average scale factor determining unit(140), and a second estimated distance calculating unit(150). [Reference numerals] (110) Image acquiring unit; (120) First estimated distance calculating unit; (130) Scale factor calculating unit; (140) Average scale factor determining unit; (150) Second estimated distance calculating unit; (160) Average scale factor verifying unit; (AA) Inputting taken images; (BB) Outputting distance between vehicles
Abstract:
본 발명은 차선검출방법 및 이를 이용한 차선이탈검출시스템에 관한 것으로, 외부입력영상을 이용하여 HSV 색상모델과 에지모델을 이용하여 차선특징을 추출하고 베이즈 결정규칙을 이용해서 특징을 결정하는 단계와, 추출된 차선특징을 이용하여 허프변환 및 필터링을 통해 차선을 검출하는 단계, 그리고 검출된 대표차선을 기준으로 차선이탈을 판단하는 단계로 구성되는 차선검출방법 및 이를 구현하는 시스템을 제공하는 것을 요지로 한다. 본 발명에 따르면, HSV 색상모델과 에지모델을 통한 특징추출과 베이즈 결정규칙에 의한 차선 특징 결정방법, 허프변환을 이용하여 효과적인 차선검출방법 및 시스템을 제공하여 차량의 주행중에 외부환경에 의한 노이즈 정보를 효율적으로 제거하여 차선의 이탈정보를 정확히 형성할 수 있는 효과가 있다. 차선검출, HSV 색상모델, 에지모델, 허프변환, 베이즈 결정규칙
Abstract:
PURPOSE: A lane detection method is provided to efficiently remove pixels including error data in features of lanes by being applied to a real-time lane deviation detection system and to create the information with high precision. CONSTITUTION: A lane detection method comprises the steps of: extracting lane features using external input images; detecting and filtering lanes using the extracted lane features; determining the lane deviation on a basis of the detected representative lanes. The color model and edge information of lanes are extracted from the input images.
Abstract:
PURPOSE: A vision based part curvature estimating method and an apparatus thereof are provided to extract an outline from a photographed image about a part having a curved surface and to estimate central point coordinates about a virtual circle having the curved surface as an arc using arbitrary coordinates of the outline corresponding to the curved surface, thereby accurately estimating an actual curvature value about the curved surface. CONSTITUTION: A part curvature estimating apparatus obtains a photographed image about a part having a curved surface (S410). The part curvature estimating apparatus extracts an outline of the part from the photographed image (S420). The part curvature estimating apparatus estimates central point coordinates about a virtual circle having the curved surface as an arc using arbitrary coordinates of the outline corresponding to the curved surface (S430). The part curvature estimating apparatus estimates curvature of the curved surface using the arbitrary coordinates and the central point coordinates (S440). [Reference numerals] (AA) Start; (BB) End; (S410) Obtain a photographed image about a part having a curved surface; (S420) Extract an outline of the part from the photographed image; (S430) Estimate central point coordinates about a virtual circle having the curved surface as an arc using arbitrary coordinates of the outline corresponding to the curved surface; (S440) Estimate curvature of the curved surface using selected arbitrary coordinates forming the outline and the central point coordinates; (S450) Compare the estimate curvature information to the basic curvature information; (S460) Correction control of the curvature