电子测量与仪器学报
電子測量與儀器學報
전자측량여의기학보
JOURNAL OF ELECTRONIC MEASUREMENT AND INSTRUMENT
2014年
10期
1140-1148
,共9页
周云鹏%朱青%王耀南%卢笑%凌志刚
週雲鵬%硃青%王耀南%盧笑%凌誌剛
주운붕%주청%왕요남%로소%릉지강
疲劳检测%Gabor滤波%梯度信息%LBP特征金字塔%模糊系统
疲勞檢測%Gabor濾波%梯度信息%LBP特徵金字塔%模糊繫統
피로검측%Gabor려파%제도신식%LBP특정금자탑%모호계통
drowsiness detection%Gabor filter%gradient information%LBP feature pyramid%fuzzy system
为了监测驾驶员的疲劳状态,提出了一种基于面部多种疲劳参数的驾驶员状态检测算法。首先利用Gabor滤波和梯度信息增强眼睛和嘴部的边缘信息以进行准确定位,然后采用一种旋转不变的LBP金字塔特征对眼睛进行特征描述,训练线性SVM分类器判别眼睛的开闭状态;并根据嘴部的张开面积及宽高比判断嘴部的开闭状态,同时通过统计眼睛在垂直方向上的运动确定头部位置的变化。最后基于眼睛和嘴部的状态、头部的位置,计算出4个能够描述驾驶员状态的疲劳参数,利用模糊系统推理得出驾驶员最终的疲劳状态。实验结果证明检测和状态判别的算法都有较高的准确率,其中眼睛状态的识别率平均在97%,嘴部状态的识别率也能达到92%;模糊系统的合理性也在实验中得以验证。
為瞭鑑測駕駛員的疲勞狀態,提齣瞭一種基于麵部多種疲勞參數的駕駛員狀態檢測算法。首先利用Gabor濾波和梯度信息增彊眼睛和嘴部的邊緣信息以進行準確定位,然後採用一種鏇轉不變的LBP金字塔特徵對眼睛進行特徵描述,訓練線性SVM分類器判彆眼睛的開閉狀態;併根據嘴部的張開麵積及寬高比判斷嘴部的開閉狀態,同時通過統計眼睛在垂直方嚮上的運動確定頭部位置的變化。最後基于眼睛和嘴部的狀態、頭部的位置,計算齣4箇能夠描述駕駛員狀態的疲勞參數,利用模糊繫統推理得齣駕駛員最終的疲勞狀態。實驗結果證明檢測和狀態判彆的算法都有較高的準確率,其中眼睛狀態的識彆率平均在97%,嘴部狀態的識彆率也能達到92%;模糊繫統的閤理性也在實驗中得以驗證。
위료감측가사원적피로상태,제출료일충기우면부다충피로삼수적가사원상태검측산법。수선이용Gabor려파화제도신식증강안정화취부적변연신식이진행준학정위,연후채용일충선전불변적LBP금자탑특정대안정진행특정묘술,훈련선성SVM분류기판별안정적개폐상태;병근거취부적장개면적급관고비판단취부적개폐상태,동시통과통계안정재수직방향상적운동학정두부위치적변화。최후기우안정화취부적상태、두부적위치,계산출4개능구묘술가사원상태적피로삼수,이용모호계통추리득출가사원최종적피로상태。실험결과증명검측화상태판별적산법도유교고적준학솔,기중안정상태적식별솔평균재97%,취부상태적식별솔야능체도92%;모호계통적합이성야재실험중득이험증。
In order to detect the drowsiness state of the driver , a detection algorithm which is based on multi-parame-ters of facial drowsiness is proposed.Firstly, the marginal information of the eyes and the mouth are enhanced by using Gabor filter and gradient information to localize them precisely , and then the state of the eyes is classified as being open and closed by a linear SVM classifier trained by an rotation-invariant LBP features of the eye image , and the state of the mouth is determined according to the opening-area and the ratio of width-height of the mouth .And meanwhile the change of the head position is confirmed by the statistics of the movement of the eyes in the vertical o-rientation.Finally, the state of the driver can be reasoned by the fuzzy system using four drowsiness ’ s coefficients calculated by the state of the eyes and the mouth , and the head position.The experiment results prove that the detec-tion and the state discrimination algorithm have higher accuracy , of which the average accuracy of the recognition of the eye state amounts to 97%, and the recognition of mouth state can also reach to 92%.The reasonability of the fuzzy system is verified in the experiment as well .