This design analyzed the shape parameters of weeds. Through the data comparison and experimental verification, it comprehensively utilized the six effective shape parameters as a neural network input feature vector, which contained the broad and long edge ratio, the leafage and circumcircle area ratio, the leafage and girth's square ratio ,the leafage and circum-rectangle area ratio, the framework and area ratio, the framework and girth ratio. In this article,neural network was trained and improved. The frequency, sample rate and the recognition correction rate for the test samples should be considered in the network structure. The experimental results showed that Neural Network Classifier was able to identify crops and weeds well. For specific plants, the recognition rate could be improved using of a specific shape features. © 2012 Springer-Verlag GmbH.
CITATION STYLE
Li, D. M., Du, B., & Zhang, L. (2012). The design of the weeds classification system based on BP neural network. In Advances in Intelligent and Soft Computing (Vol. 169 AISC, pp. 625–630). https://doi.org/10.1007/978-3-642-30223-7_99
Mendeley helps you to discover research relevant for your work.