Judgment method for atmospheric density of space

The invention discloses a judgment method for atmospheric density of space. The judgment method comprises the following specific implementation steps: a user provides an input parameter which is recorded as in; the in is subjected to normalization processing; a constant term 1 is added to a parameter sequence subjected to normalization operation; the input sequence is multiplied by the weight matrix of an input layer of a neural network so as to enable the calculation times n of a hidden layer of the neural network to be equal to 1; the constant term 1 is added to the input sequence to allow a (n) to be equal to [a(n); 1], and the transformed input parameter sequence a(n) is multiplied by LW(n), wherein a(n) is equal to LW(n)*a(n), so calculation of the hidden layer is completed; n is set to be equal to n+1, and if n is greater than 3, calculation of the hidden layer is terminated and a next step is to be conducted, or otherwise, n is made to be equal ton+1, and the previous step is repeated to complete the c