Biological sample sorting method, surface acoustic wave microfluidic chip, system, terminal, and storage medium

A biological sample sorting method, comprising:

• injecting a mixed solution containing biological samples into a surface acoustic wave microfluidic chip, and applying a first sinusoidal signal to the surface acoustic wave microfluidic chip thus forming a standing wave sound field in the surface acoustic wave microfluidic chip and making the biological samples aggregate at a node position in the standing wave sound field under the action of an acoustic radiation force;
• collecting images of the aggregated biological samples;
• performing moving target identification and tracking on the images of the biological samples using a deep learning model, performing cluster analysis and classification on a tracked moving target to obtain at least one target sample, and generating a delay enabling signal of the at least one target sample according to a moving speed of the at least one target sample;
• and
• applying a second sinusoidal signal to the surface acoustic wave microfluidic chip according to the delay enabling signal to control the at least one target sample to move thus sorting out the at least one target sample;
• wherein the deep learning model comprises a Fully Convolutional Network (FCN) network and a Siamese neural network, and wherein the operations of performing moving target identification and tracking on the images of the biological samples comprise:
• performing moving target recognition on the biological samples in the images of the biological samples by using a moving target detection algorithm based on the FCN model;
• and
• tracking a recognized moving target by using a moving target tracking algorithm and calculating a moving speed of the moving target in real time based on the Siamese neural network.