配制了与人体软组织性能相近的水凝胶人造软组织。采用由光学平台、相机及支架、相机光源等组成的图像采集设备,记录嵌入在软组织内标识物的连续位移,研究穿刺过程中软组织的变形规律。在分析标识物在X方向和Y方向上位移的基础上,基于反向传播(BP)神经网络,建立标识物在Y方向上的神经网络模型。通过与实验数据对比,神经网络模型的拟合度在95%以上,有效数据的最大相对误差控制在30%,最大绝对误差为0.8mm,能够较好地定量预测穿刺过程中软组织的变形。研究结果可有效提高软组织针穿刺靶点精度。 Formulated hydrogel artificial soft tissue with similar properties to human soft tissue. An image acquisition device composed of an optical platform, a camera and a bracket, and a camera light source is used to record the continuous displacement of the marker embedded in the soft tissue to study the deformation of the soft tissue during the puncture. Based on the analysis of the displacements of the markers in the X and Y directions, a neural network model of the markers in the Y direction is established based on the Back Propagation (BP) neural network. Compared with the experimental data, the fitting degree of neural network model is more than 95%, the maximum relative error of effective data is controlled at 30% and the maximum absolute error is 0.8mm, which can predict the deformation of soft tissue well during puncture. The research results can effectively improve the accuracy of soft tissue needle puncture target.