CAJ | 학술논문

以玉米病害图像为例,经图像预处理后,采用遗传算法从图像纹理、颜色和形状多个原始特征中优选,优化出相关信息测度、归一化蓝色颜色分量 b 值、Cb、颜色矩、病斑周长和形状因子等独立性、稳定性好及分类能力强的特征向量用于病害识别.利用SPSS软件提供的Bayes判别分析结果表明,该方法提高了病害图像识别的效率和精度.
이옥미병해도상위례,경도상예처리후,채용유전산법종도상문리、안색화형상다개원시특정중우선,우화출상관신식측도、귀일화람색안색분량 b 치、Cb、안색구、병반주장화형상인자등독립성、은정성호급분류능력강적특정향량용우병해식별.이용SPSS연건제공적Bayes판별분석결과표명,해방법제고료병해도상식별적효솔화정도.
In this article, we took the pretreated maize disease images as example, and employed genetic al-gorithms to choose approximate and effective image features, including relevant information measure, color components b and Cb, color moments, lesion perimeter, shape factor etc as recognition features from many pri-mordial features. The Bayes discriminant analysis results provided by SPSS software show that this method can improve the efficiency and accuracy of the disease image recognition.