CAJ | 학술논문

考虑算法的实时性，交通流个体识别往往采用轮廓特征来描述。由于自行车与行人的很多轮廓特征值相近，在复杂城市交通环境中，从慢行交通流中准确地识别出自行车是目前采用视频检测交通流的难点之一。通过现场视频检测获取足够的样本外轮廓数据，分析检测目标轮廓的高、宽、高宽比、面积等典型轮廓特征，发现目标轮廓高宽比特征能较好地识别出行人、自行车等慢行交通，并给出合适的取值范围。研究表明自行车高宽比能够较好地成为视频识别的依据，为自行车视频识别提供数据。
고필산법적실시성，교통류개체식별왕왕채용륜곽특정래묘술。유우자행차여행인적흔다륜곽특정치상근，재복잡성시교통배경중，종만행교통류중준학지식별출자행차시목전채용시빈검측교통류적난점지일。통과현장시빈검측획취족구적양본외륜곽수거，분석검측목표륜곽적고、관、고관비、면적등전형륜곽특정，발현목표륜곽고관비특정능교호지식별출행인、자행차등만행교통，병급출합괄적취치범위。연구표명자행차고관비능구교호지성위시빈식별적의거，위자행차시빈식별제공수거。
Consider the real‐time performance of algorithm ,contour feature is often used to describe the individual of traffic flow ;Because the contour feature of bicycles are similar to pedestrians’ ,identification of the bike from the slow traffic flow accurately is one of the difficulties of using video detection of traffic flow in the complex urban traffic environment .Based on the enough sample’s contour data w hich through video detection ,this paper mainly analyzes and detect target’s width ,height ,area and aspect ratio ,and find out that the target contour’s aspect ratio feature can be a very good recognition for pedestrians ,bicycles and other slow traffic tools .Finally ,this paper presents the suitable scope of bicycle aspect ratio .Studies have show n that the bicycle’s aspect ratio of the video can be identified as a good basis for the bicycle video recognition contribute .