This website requires JavaScript.
DOI: 10.1109/ICBAIE56435.2022.9985932

Deep Cost-sensitive Learning for Wheat Frost Detection

Shujian CaoLin CuiHaipeng Liu
Dec 2022
摘要
Frost damage is one of the main factors leading to wheat yield reduction.Therefore, the detection of wheat frost accurately and efficiently isbeneficial for growers to take corresponding measures in time to reduceeconomic loss. To detect the wheat frost, in this paper we create ahyperspectral wheat frost data set by collecting the data characterized bytemperature, wheat yield, and hyperspectral information provided by thehandheld hyperspectral spectrometer. However, due to the imbalance of data,that is, the number of healthy samples is much higher than the number of frostdamage samples, a deep learning algorithm tends to predict biasedly towards thehealthy samples resulting in model overfitting of the healthy samples.Therefore, we propose a method based on deep cost-sensitive learning, whichuses a one-dimensional convolutional neural network as the basic framework andincorporates cost-sensitive learning with fixed factors and adjustment factorsinto the loss function to train the network. Meanwhile, the accuracy and scoreare used as evaluation metrics. Experimental results show that the detectionaccuracy and the score reached 0.943 and 0.623 respectively, this demonstrationshows that this method not only ensures the overall accuracy but alsoeffectively improves the detection rate of frost samples.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答