储粮害虫图像识别中的特征提取

特征提取是储粮害虫图像识别中的重要环节，是识别系统的难点所在。针对粮虫的二值化图像提取出17个形态学特征，并进行归一化处理；把交叉验证训练模型的识别率作为储粮害虫特征提取评价准则的一个重要因子，运用蚁群优化算法从粮虫的17维形态学特征中自动提取出面积、周长等7个特征的最优特征子空间；采用支持向量机分类器对9类粮虫进行分类，识别率达到95%以上，证实了基于蚁群优化算法的粮虫特征提取的可行性。

Feature extraction for the stored-grain insect detection system based on image recognition technology

The feature extraction is a very important and difficult part for the stored-grain insect detection system based on image recognition technology. The seventeen morphological features were extracted and normalized from the binary grain-insect images. The ant colony optimization algorithm was applied to the feature extraction of the stored-grain insects, and the recognition accuracy of the z-fold cross-validation training model was acted as an important factor for the evaluation principle of the feature extraction. The algorithm extracted seven features that were composed of the optimal feature space from the 17 morphological features, such as area and perimeter. Finally, the nine species of the stored-grain insects were recognized by the support vector machine classifier, and the correct identification ratio was over 95%. The experimental results show that the feature extraction of the stored-grain insect based on ant colony optimization algorithm is practical and feasible.