When star sensors operate under near -Earth daytime conditions, the intense background radiation from the sky severely interferes with the energy of the star points in the imagery, resulting in a low signal-to-noise ratio (SNR) for the star points. This low SNR hinders target extraction and centroid positioning, thereby affecting the normal attitude measurement of star sensors. Addressing the challenge of attitude measurement under daytime conditions, this study first analyzes the mapping relationship of pixel positions in consecutive frames of star sensor imagery. A star image superposition algorithm based on attitude -related frames is proposed. On this foundation, an attitude measurement method based on star image superposition is employed for measuring the attitude of daytime star sensors. Furthermore, a fitting algorithm for the solar centroid is introduced, and a coarse measurement method based on solar position is applied to determine the optical axis orientation of daytime star sensors, enhancing their robustness in daylight conditions. The algorithm proposed in this study is validated through experiments. The results demonstrate that the multimodal attitude measurement method not only effectively improves the SNR of near -Earth daytime star sensor imagery through star image superposition, ensuring the accuracy of attitude measurements, but also ensures the robustness of attitude measurements through solar centroid fitting.

Under the dynamic working conditions of a star sensor, motion blur of the star will appear due to its energy dispersion during imaging, leading to the degradation of the star centroid accuracy and attitude accuracy of the star sensor. To address this, a restoration method of a blurred star image for a star sensor under dynamic conditions is presented in this paper. First, a kinematic model of the star centroid and the degradation function of blurred star image under different conditions are analyzed. Then, an improved curvature filtering method based on energy function is proposed to remove the noise and improve the signal-to-noise ratio of the star image. Finally, the Richardson Lucy algorithm is used and the termination condition of the iterative equation is established by using the star centroid coordinates in three consecutive frames of restored images to ensure the restoration effect of the blurred star image and the accuracy of the star centroid coordinates. Under the dynamic condition of 0~4°/s, the proposed algorithm can effectively improve the signal-to-noise ratio of a blurred star image and maintain an error of the star centroid coordinates that is less than 0.1 pixels, which meets the requirement for high centroid accuracy.

The tracking performance of star sensor degrades seriously under dynamic conditions. To improve the tracking accuracy and efficiency, an attitude tracking method based on unscented Kalman filter (UKF) and singular value decomposition (SVD) is proposed in this paper. The star sensor is modeled as a nonlinear stochastic system, the state of which is attitude quaternion. The quaternion can be estimated by UKF, then the predicted attitude and corresponding star positions are obtained. To ensure the stability of attitude tracking, SVD is applied to obtain the sigma points in UKF continuously. The experimental results indicate that the proposed method yields high accuracy and efficiency in attitude tracking. This method provides a practical approach to ensure the tracking performance of star sensor under dynamic conditions.

The detection and recognition of arrow markings is a basic task of autonomous driving. To achieve all-day detection and recognition of arrow markings in complex environment, we propose a hybrid model by exploiting the advantages of biologically visual perceptual model and discriminative model. Firstly, the arrow markings are extracted from the complex background in the region of interest (ROI) by the biologically visual perceptual model using the frequency-tuned (FT) algorithm. Then candidates for road markings are detected as maximally stable extremal regions (MSER). In recognition stage, biologically visual perceptual model calculates the sparse solution of arrow markings using sparse learning theory. Finally, discriminative model uses the Adaptive Boosting (AdaBoost) classifier trained by sparse solution to classify arrow markings. Experimental results show that the hybrid model achieves detection and recognition of arrow markings in complex road conditions with the precision, recall, and F-measure being 0.966, 0.88, and 0.92, respectively. The hybrid model is robust and has some advantages compared with other state-of-the-art methods. The hybrid model proposed in this paper has important theoretical significance and practical value for all-day detection and recognition in complex environment.

Generative adversarial networks had shown promising potential in conditional image generation. It seemed that the GANs were particularly suitable for use in image super-resolution reconstruction. However, there was a shortcoming of excessive smoothness and lack of high frequency detail information for the reconstructed SR images by using GANs. Aiming at resolving the problem that the method of single image super-resolution reconstruction ignored the spatio-temporal relationship between image frames, a method of multiframe infrared image super-resolution reconstruction based on generative adversarial networks (M-GANs) was proposed in this paper. Firstly, motion compensation was proposed for registration low resolution image frames; Secondly, a weight representation convolutional layer was performed to calculate the weight transfer; Finally, the generative adversarial network was used to reconstruct the high resolution image. Experimental results demonstrate that the proposed method surpass current state-of-the-art performance of both subjective and objective evaluation.

To solve the space coherence of four inference fringe images by dynamic interferometer with common light paths, we propose a new notion of inference fringe image registration. We detach the conjunction between the inference fringe image and the image registration 7 and erect the equipment for inference fringe image registration 7 avoiding the confusion between the inference fringe and the cross wire which can influence the measure precision of dynamic interferometer. First, we realize the physical registration of four CMOS cameras using the equipment for inference fringe image registration. Then, we go along the image registration for the four cross wire images of the depict board by cross wire extraction 7 intersection point caculation and rotution amount caculation using total least square method 7 and realize the parallelism between the pix and pix of the dynamic interferometer with common light paths. Last, experimental results show that the proposed registration algorithm can improve the accuracy of registration 7 which is superior to that of the method of bary center. The method proposed in this paper can achieve the cross-correlation value of over 96%.

为实现远距离、高可靠性传输,并减小复杂度,对Camera Link Full接口数据的HD-SDI传输显示进行了深入研究.采用FPGA作为核心处理器,考虑相机输出具有多种帧频,采取帧频检测及充分降频策略,并通过3个SRAM进行缓存以实现帧频转换,以满足HD-SDI帧频25Hz的要求.考虑到SRAM数据宽度,采取FIFO行缓存策略将Camera LinkFull80输出的10 tap、80 bits图像数据转换成单通道的8 bits图像数据.最后,完成系统设计并进行实验验证.实验结果表明:系统实现了图像数据从50 Hz、100 Hz、500 Hz等多种帧频的Camera Link Full80到25帧HD-SDI接口1080i的格式转换及实时显示,且图像层次丰富,无失真.

为了提高光电经纬仪测角精度,实现经纬仪对目标的精确空间定位,本文设计了外场星体标校系统.介绍了系统组成、工作原理及试验结果分析.利用xxxx弹道相机作为载体,进行了设计、加工、运输,在外场进行了试验.试验结果表明:该弹道相机可实现精确定位,精度优于10″,满足指标要求.

Three dimensional (3D) object recognition was researched under multi-view points. For the shortages of traditional signal feature description for 3D object recognition under multi-view points, a new recognition algorithm fusing multiple features was proposed. Firstly, the object corners were extracted by using the correlation matrixes of anisotropic Gaussian directional derivatives, the particular corners were selected by the skeleton constraint, and the normalized distance between particular corners and the object centroid was taken as the corner descriptor. Then, the geometric moment invariants, affine moment invariants, and the Fourier descriptor of object boundary were extracted, respectively, and the scatter matrixes within and between classes for the four features were calculated. By taking the trace of sample scatter matrix as the weight, the four features were fused. Furthermore, the Independent Component Analysis (ICA) was carried out on the fused vector to obtain independent features. Finally, a Support Vector Machine (SVM) was adopted to complete the whole classification of the experiments. Experimental results show that the recognition accuracy of the proposed approach is higher than that of the signal feature approach by 10% averagely and that in the small training sample (10% of the total samples) condition still achieves more than 80%.It concludes that proposed algorithm meets the demand of theodolites for real-time object recognition.