Click on “Download PDF” for the PDF version or on the title for the HTML version.
If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.
Corn Seeding Monitoring System Based on Image Processing
Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.orgCitation: 2018 ASABE Annual International Meeting 1800354.(doi:10.13031/aim.201800354)
Authors: Baosheng Li, Xingxing Liu, Tong Zhang, Fang Tian, Yu Tan, Shixiong Li
Keywords: precise seeding, image processing, monitoring system, agriculture.
During the operation of precise seeding, a high-speed camera installed under the seeding device collected the images of the process of seeds falling in real time and transferred the collected images to the embedded detection system based on a STM32 microcontroller. The system used the algorithm of Artificial Neural Network to analyze the images and detect whether the seeds are correctly sowed. Convolutional neural network was used as the input layer to reduce the data and extract useful features to form the feature map. Moreover, a fully connected network was used in the middle layer to combine and summarize the features from the former layer. SoftMax classifier was used as the output layer to give the prediction results, i.e., whether seeds were correctly sowed. The training of the model of the neural network was completed in a high-performance workstation. The mature network was transplanted to the embedded control system of the microcontroller for real-time detection of seeds.
The test indicated that the whole system run smoothly and accurately during monitoring. It is convenient to install and can be widely used in corn precision seeders and intelligent agricultural production.(Download PDF) (Export to EndNotes)