A Model for Identifying Soybean Growth Periods Based on Multi-Source Sensors and Improved Convolutional Neural Network

The identification of soybean growth periods is the key to timely take field management measures, which plays an important role in improving yield. In order realize discrimination under complex environments quickly and accurately, a model for identifying based on multi-source sensors improved convolutional neural network was proposed. AlexNet structure by adjusting number fully connected layer 1 2 neurons 1024 256. optimized through hyperparameters combination experiment classification different types image datasets. emergence (VE), cotyledon (VC), first node (V1) stages achieved. experimental results showed that after layers, average accuracy 99.58%, loss 0.0132, running time 0.41 s/step optimal hyperparameters. At around 20 iterations, performances began converge were all superior baseline model. Field validation trials conducted applying model, 90.81% VE, 91.82% VC, 92.56% V1, with 91.73%, single recognition about 21.9 ms. It can meet demand smart phone unmanned aerial vehicle (UAV) remote sensing, provide technical support resolutions from sensors.