CAJ | 학술논문

　　为了提高植物叶片识别与分类的正确率，提出了基于SVM 的识别模型和方法；对叶片图像预处理后，提取并优选10个叶片形状特征参数，用SVM 法进行训练建模并识别。实验结果表明，用线性核函数的 SVM 对木瓜、女贞、三角枫和五角枫4种植物叶片识别的平均准确率在95．8％以上，优于神经网络和 Fisher 判别法，为鉴定植物种类提供了一种快速有效的方法。
　　위료제고식물협편식별여분류적정학솔，제출료기우SVM 적식별모형화방법；대협편도상예처리후，제취병우선10개협편형상특정삼수，용SVM 법진행훈련건모병식별。실험결과표명，용선성핵함수적 SVM 대목과、녀정、삼각풍화오각풍4충식물협편식별적평균준학솔재95．8％이상，우우신경망락화 Fisher 판별법，위감정식물충류제공료일충쾌속유효적방법。
To improve the correct rate of identification and classification of plant leaves , a model and method based on SVM was proposed.After pre-processing the leaves image, extraction,and preferably 10 leaf shape characteristic parame-ters was extracted,then use SVM method to train model and identify leaves .The results show that the linear kernel SVM has average accuracy above 95.8% for four plant leaves of papaya , privet, trident maple, acer mono,which is better than neural network and Fisher discriminant method .This research provides a fast and effective method for the identifica-tion of plant species.