Polarization detection of space targets is one of the most important research directions in the field of space target recognition. In view of the fact that there are problems such as strong background noise and inconspicuous details of contour features in the polarization image of space targets, an image denoising and enhancement strategy is proposed. To solve the problem of high intensity of Gaussian noise in degree of polarization (DoP) images, a denoising method named adaptive noise template prediction (ANTP) is proposed to eliminate the noise. Compared to the existing methods, the ANTP algorithm performs better at reducing noise and improving image quality. Aiming at the difficulty of separating the background noise from angle of polarization (AoP) images, a denoising method named gray analysis of local area (GALA) is proposed. Compared to traditional methods, the GALA algorithm can effectively extract the contour features of targets and improve the contrast of AoP images. An image fusion method based on discrete cosine transform and local spatial frequency (LSF) is used to fuse the denoised DoP image and AoP image. The experimental results of the simulated and real space target polarization detection confirm the effectiveness of our proposed strategy.

A low threshold current density of 2 μm InGaAsSb/AlGaAsSb muti-quantum well (MQW) laser diode with carrier blocking layer (CBL) is demonstrated by simulation and fabrication. The carrier leakage is found to be theoretically suppressed for the devices with CBL. All the laser wafers are grown with a solid source Molecular Beam Epitaxy(MBE) System. Experimental results reveal the samples with CBL exhibits ultra-low threshold current densities of 142 A/cm² and high slope efficiency of 0.158 W/A, which is better than 215 A/cm² and 0.122 W/A achieved in the conventional InGaAsSb/AlGaAsSb LDs at room temperature. This improvement in device performance comes from meticulously designing the carrier blocking layers to increase carrier confinement and injection efficiency.

As one of the major error sources, satellite signature effect should be reduced or even erased from the distribution of the post-fit residuals to improve the ranging precision. A simulation of satellite signature effect removal process for normal point algorithm is conducted based on a revised model of satellite response, which fully considers the structural and distribution characteristics of retroreflectors. In order to eliminate both long-term and short-term satellite signature effect, a clipping method for SLR data processing is proposed by defining the clipping location as 5.6 mm away from the mean value of the long-term fit residuals to select effective returns for normal points. The results indicate that, compared to normal points algorithm, the RMS per NP of LAGEOS-1 observation data processed by the clipping method is reduced from 62.90 ± 9.9 mm to 56.07 ± 4.69 mm, and the stability of RMS is improved 53%. This study improves the satellite signature effect model and simulates the fluctuation of normal points caused by satellite signature effect for the first time. The new method based on the simulation of satellite signature effect has stronger robustness and applicability, which can further minimize the influence of satellite signature effect on the SLR production and significantly improve the data property.