Method for rapidly measuring root moisture of masson pine sapling based on weighted autocoder
An automatic encoder and measurement method technology, applied in the field of chemical substance content measurement in seedlings, can solve the problem of inability to effectively extract high-order features of spectral data, and achieve the effects of accurate reconstruction, accurate and rapid measurement, and improved efficiency and precision.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0028] Such as figure 1 As shown, it is a typical autoencoder structure diagram, which consists of two parts: an encoder and a decoder, which essentially perform some transformation on the input signal. This structure converts the input layer data x to the middle layer (hidden layer) h, and then to the output layer. Each node in the graph represents a dimension of the data, and the transformation between each two layers is "linear transformation" + "nonlinear activation function". Suppose the input data is x=[x (1) , x (2) ,...x (d) ]', d is the dimension of the input data. The encoder passes the function h=f(x)=s f (Wx+b) maps the input x to the hidden layer where W is d h *d weight matrix, b is the bias vector, s f is a nonlinear activation function, d h is the dimensionality of the hidden layer vector h. Then, the decoder passes the function Map the hidden layer representation h to the output layer in is d*d h weight matrix, is the bias vector of the out...
Embodiment 2
[0036] In order to further highlight the advantages of the present invention, a batch of masson pine seedlings have been tested, the method is the same as in embodiment 1, 118 samples are divided into 88 calibration data sets and 30 prediction data sets, and the experimental results are compared with the partial minimum two Multiplicative regression, support vector regression, stacked autoencoder combined with artificial neural network, stacked autoencoder combined with support vector regression, and weighted stacked autoencoder combined with artificial neural network methods were compared, and the results are shown as Figure 4 , in the figure, a partial least squares regression, b support vector regression, c stacked autoencoder combined with artificial neural network, d stacked autoencoder combined with support vector regression, e weighted stacked autoencoder combined with artificial neural network, f weighted stacked Autoencoders combined with support vector regression.
...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More