The invention discloses a
radar moving target multi-frame joint detection method based on a graph space-time network. The method mainly solves the problem that in the prior art, the
false alarm rate of moving target detection of a single-channel
system is high. According to the scheme, the method comprises the steps of obtaining a sub-aperture distance Doppler spectrum; constructing a sub-residualnetwork and a sub-graph space-time network, and forming a neural
network model for moving target detection by using the sub-residual network and the sub-graph space-time network; performing regionaltarget detection by using the sub residual network, outputting a preliminary detection probability graph, and calculating
cross entropy loss; performing spatial-temporal
feature extraction and fusionby using a sub-graph spatial-temporal network, outputting a final detection probability graph of an intermediate frame moving target, and calculating a
mean square error; and taking the sum of the
cross entropy loss and the
mean square error as a total cost function, training the neural network until the total cost function is converged to obtain a trained neural network, inputting
test data intothe trained neural network, judging an output threshold value of the trained neural network, and suppressing a non-maximum value to obtain a moving target detection result of an intermediate frame. According to the invention, the
false alarm rate is reduced, and reliable moving target detection can be realized.