A platform for research: civil engineering, architecture and urbanism
The invention discloses a parking space line identification method based on vision, and relates to the field of parking space identification of automatic parking. The method comprises the steps of obtaining a look-around image, obtaining parking space features by performing feature extraction on the look-around image, identifying slot points and parking space types, and calculating center coordinates of slot point positioning frames; if identifying that a plurality of parking spaces exist in the look-around image, performing pairing on the slot points in pairs according to the fact that the center coordinates of the two slot point positioning frames of the same parking space are located in a parking space positioning frame; calculating coordinates of the remaining two angular points of the same parking space according to the parking space type and the matched center coordinates of the two slot point positioning frames; processing the image according to the coordinates of the four angular points of the parking space to obtain a parking space image, extracting image features, and judging whether an obstacle exists in the parking space or not; and if no obstacle exists in the parking space, the parking space can be parked, and outputting the coordinates of the four angular points of the parking space capable of being parked. The parking space can be accurately detected in the complex parking space and the complex illumination environment.
本发明公开了一种基于视觉的车位线识别方法,涉及自动泊车的车位识别领域,包括:获取环视图像,通过对环视图像进行特征提取,得到车位特征,识别槽位点以及车位类型,计算槽位点定位框的中心坐标;如果识别到环视图像中存在多个车位,根据同一个车位的两个槽位点定位框的中心坐标位于车位定位框内,对槽位点进行两两配对;根据车位类型,以及匹配好的两个槽位点定位框的中心坐标,计算同一车位剩余两个角点坐标;根据车位四角点坐标对图像进行处理,得到车位图像,提取图像特征,判断车位内是否有障碍物;若车位内无障碍物,车位可停靠,输出可停靠车位四角点坐标。本发明在面对复杂停车位,复杂光照环境时,能够准确检测停车位。
The invention discloses a parking space line identification method based on vision, and relates to the field of parking space identification of automatic parking. The method comprises the steps of obtaining a look-around image, obtaining parking space features by performing feature extraction on the look-around image, identifying slot points and parking space types, and calculating center coordinates of slot point positioning frames; if identifying that a plurality of parking spaces exist in the look-around image, performing pairing on the slot points in pairs according to the fact that the center coordinates of the two slot point positioning frames of the same parking space are located in a parking space positioning frame; calculating coordinates of the remaining two angular points of the same parking space according to the parking space type and the matched center coordinates of the two slot point positioning frames; processing the image according to the coordinates of the four angular points of the parking space to obtain a parking space image, extracting image features, and judging whether an obstacle exists in the parking space or not; and if no obstacle exists in the parking space, the parking space can be parked, and outputting the coordinates of the four angular points of the parking space capable of being parked. The parking space can be accurately detected in the complex parking space and the complex illumination environment.
本发明公开了一种基于视觉的车位线识别方法,涉及自动泊车的车位识别领域,包括:获取环视图像,通过对环视图像进行特征提取,得到车位特征,识别槽位点以及车位类型,计算槽位点定位框的中心坐标;如果识别到环视图像中存在多个车位,根据同一个车位的两个槽位点定位框的中心坐标位于车位定位框内,对槽位点进行两两配对;根据车位类型,以及匹配好的两个槽位点定位框的中心坐标,计算同一车位剩余两个角点坐标;根据车位四角点坐标对图像进行处理,得到车位图像,提取图像特征,判断车位内是否有障碍物;若车位内无障碍物,车位可停靠,输出可停靠车位四角点坐标。本发明在面对复杂停车位,复杂光照环境时,能够准确检测停车位。
Parking space line identification method based on vision
一种基于视觉的车位线识别方法
QIN YI (author)
2021-06-18
Patent
Electronic Resource
Chinese
Research on Parking Space Status Recognition Method Based on Computer Vision
DOAJ | 2022
|Charging pile parking space automatic identification parking lock system
European Patent Office | 2022
|Parking space line arrangement device for arranging automatic parking test
European Patent Office | 2024
|Overall parking space sharing method based on parking robot
European Patent Office | 2021
|