图像识别处理技术在农业工程中的应用

Application of image recognition and processing technology in agricultural engineering

针对图像识别技术在农业工程中的应用研究了一种苹果图像识别技术,为实现苹果智能采摘提供理论依据。通过图像识别技术,对苹果图像进行预处理,即对图像进行分割与特征提取。使用反向传播神经网络(BPNN)对得到的特征进行学习,得到了能够准确识别苹果图像的识别模型。为了提高BPNN模型的性能,使用遗传算法对神经网络模型中的阈值和权值进行优化。通过实例对模型的性能进行了验证,结果表明:相较常规BPNN模型,提出的研究方法具有更高的识别准确率。

Application of image recognition technology in agricultural engineering is researched,main research is apple image recognition technology,it provides a theoretical basis for realization of apple intelligent picking. Image recognition technology is used for apple image preprocessing,that is image segmentation and feature extraction. Use back propagation neural network（ BPNN） to get the feature to learn,so as to obtain recognition model which is able to accurately identify apple image. In order to improve performance of BPNN model,use genetic algorithm to optimize the threshold and weight of NN model. Through the example,performance of apple fruit recognition model is verified. The results show that compared with the conventional BPNN model,this research method has higher recognition accuracy.