结合椭圆定位方法，分别采用了线型、十字型和时钟型超声压电晶片阵列对金属薄板中缺陷进行成像识别。根据Lamb波在薄板中的传播分析，选取激励信号及其参数，构建超声信号检测模型。依据椭圆轨迹原理进行成像，判断薄板中有无缺陷存在。通过压电陶瓷晶片加载脉冲激励信号，分别对三种情况进行数据处理并成像。结果证明十字型和时钟型阵列能有效地对缺陷成像，并避免了成像过程中的虚像。

In the paper, ultrasonic linear array, cross array and clock-type array are used to image for defects of a thin plate with elliptical location method. According to Lamb wave propagation in the plate, the excitation signal and its parameters are selected and a model of ultrasonic signal detection is built up. Imaging is made based on the principle of elliptical trajectory and whether defects exist in the plate was determined. Excitation signal is generated by the piezoelectric ceramic wafer. Above-mentioned three kinds of arrays are used to acquire signals, which are then processed and imaged. The results show that it is effective to use cross array and clock-type array to do defect imaging and remove the effect of ghost generated in the process of imaging.