ing at the phenomena of abnormal response and dark signal tailing caused by crosstalk of a linear array detector, we analyze the mechanism of crosstalk generation, establish the RC model which can reproduce the crosstalk image signal waveform. On this basis, we further propose a crosstalk image restoration method. This method is based on the crosstalk RC model and the objective is to recover the normal signal response and eliminate the dark signal tailing. The optimization objective function is established. Based on the target response curves of different frequencies, the model parameters are iterated to obtain the corresponding model parameters when the comprehensive effect of restoration is optimal for each image frequency. After the parameters of the crosstalk model are obtained, the corresponding restoration function is calculated, and the image is restored by the operations in the frequency domain. The infrared images of the linear array detector scanning camera acquired in the laboratory are restored and the results show that the proposed method can effectively restore the normal response of different targets under different image frequencies, reduce the effect of tailing dark signals, and improve the image quality.

ing at the problems that the measurement range of a monocular light pen three-dimensional measurement system is small and this system can not realize the full space measurement, a full space light pen monocular vision measurement method based on precision rotary platform is proposed. First, the camera is fixed on the precision rotary platform, and the calibration plate is photographed at different angles to obtain the position of the camera optical center in the calibration plate coordinate system under different angles. Second, through the plane fitting of principal component analysis (PCA), the plane where the camera optical center is located and the rotation axis direction vector of the camera rotation motion are obtained, and the position of the camera rotation axis is obtained by using the spatial least square circle fitting. Third, with the help of the precision rotary platform reading and Rodriguez formula, the rotation angle of the turntable is transformed into the rotation matrix and translation vector of the camera. Finally, the measurement data of the camera at different positions after rotating a certain angle are converted to the same rotary platform coordinate system by using the calculated transformation matrix to realize the full space measurement. Experimental results show that in the measuring system using the same light pen, the measurement method in this paper can almost achieve the same precision as that of the traditional monocular vision light pen measurement system, achieve 360° full space measurement, and greatly expand the scope of applications of the monocular light pen.

ing at the real-time request of the infrared detection system for target detection, we propose a method for fast detection of infrared targets based on key points. Taking the target center as the key point of target detection, we first design a lightweight feature extraction network. Then, we design a corresponding feature fusion network using the spatial and semantic information of features at different levels combined with the characteristics of small infrared targets. Finally, the prediction of target category, location and size is realized. The model is comparatively tested on the self-built aerial infrared target dataset. Compared with the classic detection models such as YOLOv3, the detection speed is greatly improved and the detection accuracy is only slightly reduced. Compared with the same type of fast detection model, Tiny-YOLOv3, the detection accuracy increases by 8.9% and the detection speed running on the central processing unit (CPU)increases by 13.9 ms/frame under the condition that the model size is compressed to 23.39% of Tiny-YOLOv3's size. The detection performance is significantly improved and the effectiveness of the method is confirmed.