论文部分内容阅读
利用计算机图像处理技术和遗传神经网络技术 ,建立了一个多层前馈神经网络 ,实现了大田玉米和背景图像的正确识别 ,并且通过获取玉米叶的色度直方图提取了玉米叶表面颜色特征 ,进而求得了玉米叶色的测定值。实验结果表明 ,玉米叶色值自动测定系统 ,识别玉米的准确率可达 91.6 % ,可以有效地测定玉米的叶色。该研究为实现大田玉米的化肥精确施用提供了理论依据。
A multi-layer feedforward neural network was established by computer image processing and genetic neural network technology to realize the correct identification of field corn and background image. The color features of corn leaf surface were extracted by obtaining the colorimetric histogram of corn leaf, Further, the measured value of corn leaf color is obtained. The experimental results show that the corn leaf color value can be measured automatically, the accuracy rate of identifying corn can reach 91.6%, and the leaf color of corn can be effectively measured. This study provides a theoretical basis for the accurate application of fertilizers in field corn.