四川兵工学报
四川兵工學報
사천병공학보
SICHUAN ORDNANCE JOURNAL
2014年
9期
69-72,91
,共5页
明俊峰%汪仁煌%岳洪伟%唐苏湘
明俊峰%汪仁煌%嶽洪偉%唐囌湘
명준봉%왕인황%악홍위%당소상
羽毛缺陷%几何特征%伪Zernike矩%SVM分类器
羽毛缺陷%幾何特徵%偽Zernike矩%SVM分類器
우모결함%궤하특정%위Zernike구%SVM분류기
feather defects%geometric characteristics%pseudo-Zernike moment%SVM classifier
针对拍照角度影响羽毛缺陷识别精度的问题，将不变矩的思想引入到羽毛缺陷识别领域。提出利用伪Zernike矩描述羽毛缺陷的几何特征，用支持向量机（SVM）进行分类识别的方法。模拟实验表明算法能够较好地容忍图像平移、旋转和尺度变化，具有识别速度快、识别精度高的特点，并且算法对噪声、光线变化适应性好。
針對拍照角度影響羽毛缺陷識彆精度的問題，將不變矩的思想引入到羽毛缺陷識彆領域。提齣利用偽Zernike矩描述羽毛缺陷的幾何特徵，用支持嚮量機（SVM）進行分類識彆的方法。模擬實驗錶明算法能夠較好地容忍圖像平移、鏇轉和呎度變化，具有識彆速度快、識彆精度高的特點，併且算法對譟聲、光線變化適應性好。
침대박조각도영향우모결함식별정도적문제，장불변구적사상인입도우모결함식별영역。제출이용위Zernike구묘술우모결함적궤하특정，용지지향량궤（SVM）진행분류식별적방법。모의실험표명산법능구교호지용인도상평이、선전화척도변화，구유식별속도쾌、식별정도고적특점，병차산법대조성、광선변화괄응성호。
Aimed at the recognition accuracy problem of feather defects caused by the impacts of camera angle ,the idea of invariant moment is introduced into feather defect recognition.To identify the feather de-fects,the method based on support vector machine (SVM),by extracting feature vector with Pseudo-Zernike moment,was proposed.Simulation experiments show that the algorithm can better tolerate the im-age translation,rotation and scaling transformation with characteristics of fast recognition speed and high recognition accuracy,and it is also good adaptability for noise and light changes.
