合成孔径雷达(SAR)由于其良好的特性被广泛应用于高分辨成像,但成像所需的庞大数据导致其难以在资源受限的平台推广应用。单比特SAR 通过将回波采样点表征为1 比特二进制数字信号,可以达到降低数据量、缓解平台负担的目的,但二值数据跳变产生的高阶谐波将导致成像质量下降。为提升单比特SAR 成像质量,提出基于卷积去量化(CDQOB)网络的无人机载条带SAR 成像方法,通过单比特子孔径数据实现运动误差估计与智能化距离-多普勒二维谱重构,进而实现低数据量下的高质量条带SAR 成像。通过实测数据的处理分析,验证了所提单比特成像方法的有效性。

Synthetic aperture radar (SAR) has been widely applied for high-resolution imaging due to its excellent characteristics. However, the large amount of data required for imaging makes it difficult to apply on resource-constrained platforms. One-bit SAR reduces data volume and alleviates the platform burden by representing echo samples as one-bit binary digital signals. Nevertheless, the high-order harmonics generated by the binary data transitions result in a degradation in imaging quality. To improve the imaging quality of one-bit SAR, this paper proposes a UAV-borne stripmap SAR imaging method based on the convolutional de-quantization onebit (CDQOB) network. The proposed method utilizes one-bit sub-aperture data to achieve motion error estimation and intelligent Range-Doppler two-dimensional spectrum reconstruction, thereby achieving high-quality stripmap SAR imaging with reduced data volume. The effectiveness of the proposed one-bit imaging method has been validated through the analysis of measured data.

TN957. 52

国家自然科学基金资助项目(62171293,62431021); 深圳市基金资助项目(JCYJ20230808105359045);装备预研教育部联合基金资助项目(8091B032224);国家杰出青年科学基金资助项目(61925108);国家自然科学基金国际合作与交流重点项目(62220106009)