10results about How to "Reduce ghosting" patented technology

The invention relates to a cross-color gamut phase-guided double-attention fusion HDR imaging method. The method comprises the following steps: step (1), carrying out data preprocessing of denoising and cleaning on an original multi-exposure image sequence; step (2), obtaining standardized image data; (3) mapping a nonlinear pixel value into a linear physical illumination value for the standardized image data through a reverse application camera response curve, and outputting a linearized multi-frame image; step (4), splicing the linearized multi-frame images on a channel dimension, and constructing a unified input tensor containing complete dynamic range information; (5) performing convolution operation on the input tensor, and extracting shallow feature representation of local textures and edges of the image; extracting shallow layer features; and step (6), introducing a double attention mechanism to perform interaction and aggregation of global features, and outputting a fused high dynamic range feature map. The method can stably work in a dynamic scene and under an extreme exposure condition, and promotes the practical application of an HDR imaging technology.

The application discloses a height measuring system and a photoetching machine device thereof, and relates to the field of photoetching machine wafer height and surface type detection. The height measuring system comprises a grating piece, an imaging assembly, a filtering assembly and a sensing piece. The grating piece is used for passing a to-be-detected light source. The to-be-detected light source forms diffraction beams with different diffraction orders which are distributed in sequence through the grating piece. The imaging assembly is arranged on one side of the grating piece where the diffraction beams are formed. The filtering assembly is arranged on one side of the imaging assembly close to the grating piece or in the light path in the imaging assembly. The sensing piece is used for receiving the diffraction beams filtered by the filtering assembly. The application diffracts the light beams through the grating piece, and realizes imaging through cooperation of the imaging assembly and the sensing piece. The filtering assembly filters the diffraction light of a certain order in the diffraction beams, thereby reducing the defocus and ghosting phenomenon, and improving the measurement precision.

This invention relates to a method and system for constructing a dynamic 3D Gaussian map based on semantic enhancement, belonging to the field of computer vision and robot simultaneous localization and mapping technology. The method utilizes a keyframe scheduler to filter keyframes, performs semantic detection on the keyframes based on a semantic segmentation network to generate an initial semantic mask, and performs geometric verification and correction on the mask using depth reprojection residuals. The corrected mask is propagated to adjacent ordinary frames using camera pose-based reprojection, and temporal voting fusion is performed to obtain a temporally consistent semantic mask sequence. A 3D Gaussian sphere map is initialized and updated in static regions, and semantically guided adaptive density control is performed in dynamic regions to suppress dynamic artifacts. Lightweight map management is then performed based on information entropy scoring. Finally, under photometric error and geometric constraints, joint nonlinear optimization of camera pose and Gaussian sphere parameters is performed and iteratively updated in a closed loop.

The present invention provides an image sensor comprising a photosensing layer for receiving incident light in a wavelength range, detecting the incident light, and generating an optical signal representing the incident light, wherein the photosensing layer has a first surface and a second surface opposite to the first surface; a redistribution layer RDL disposed on the first surface of the photosensing layer, the redistribution layer including a plurality of conductive traces for receiving an electrical signal representing the optical signal; and a protective layer disposed on the redistribution layer and opposite to the photosensing layer, the protective layer being configured to reflect incident light passing through the photosensing layer and the redistribution layer, and to direct the reflected light to the second surface of the photosensing layer and away from the second surface, wherein the protective layer is opaque and reflective to light in the wavelength range.

This invention provides a method for image stitching using line segment projection point-line fusion in sparse seascape texture scenes, relating to the field of image processing technology. The method includes the following steps: S1, acquiring the image to be stitched; S2, performing cylindrical projection processing on the image to be stitched to obtain a cylindrical projection image; S3, extracting point features from the cylindrical projection image and performing initial matching, estimating an initial homography matrix based on the matched point pairs; S4, extracting line segment features from the cylindrical projection image, calculating line segment matching scores, and determining matching line segments based on the line segment matching scores; S5, fusing the virtual matched point pairs with the point feature matching point pairs to obtain an optimized homography matrix; S6, completing image registration based on the optimized homography matrix, constructing a total cost function for the stitching seam in the overlapping region, and obtaining the stitched image. This invention can improve the accuracy and robustness of image stitching in sparse seascape texture scenes.

The application discloses a lens assembly, a display module and an electronic device. The lens assembly comprises a first lens and a semi-transparent semi-reflective film. The first lens comprises a first surface and a second surface opposite to each other, and the semi-transparent semi-reflective film is arranged on the first surface. The semi-transparent semi-reflective film comprises a modulation film layer and a dielectric film layer. The modulation film layer is arranged on the first surface of the first lens, and the dielectric film layer is arranged on the side of the modulation film layer away from the first lens. The dielectric film layer and the modulation film layer are used to cooperate to adjust the refractive index of the semi-transparent semi-reflective film, so that the circularly polarized light incident to the semi-transparent semi-reflective film at different incident angles remains circularly polarized light after entering the first lens. The lens assembly, the display module and the electronic device of the application can maintain the circular polarization state of the circularly polarized light passing through the semi-transparent semi-reflective film after the circularly polarized light is incident, so that the circular polarization degree of the circularly polarized light passing through the semi-transparent semi-reflective film is high, the influence of stray light on subsequent light propagation is reduced, thereby reducing ghosting and improving imaging effect.

This invention discloses a video frame interpolation method and system based on edge computing, belonging to the field of intelligent video processing technology. The method includes the following steps: acquiring data and generating video frame sequences, frame timestamps, and edge state tables; extracting change features to generate interpolated frame segments and segment complexity tables; inputting the segment data into an improved PWC-Net network to generate optical flow tables, residual region tables, and candidate frames; calculating the interpolation load to generate a computational cost table; calculating output interval deviation and node switching delay to generate a temporal sequence connection weight table; constructing a DAG graph of the interpolation task and inputting it into the HEFT algorithm to generate a task order and node allocation table; and performing candidate frame correction, splicing, and encoding to generate an interpolated video stream. This invention addresses the problems of computational redundancy in video frame interpolation under fluctuating edge node computing power, unstable interpolation in complex motion regions, and discontinuous temporal sequence connection between adjacent segments by introducing an improved PWC-Net network and the HEFT algorithm.

The invention discloses an X-ray image splicing method and device and an X-ray photography system. The X-ray image splicing method comprises the steps that multiple to-be-spliced X-ray images of a splicing body position are obtained, a candidate splicing area of at least one X-ray image is obtained, the multiple X-ray images are displayed on a splicing planning interface, and the candidate splicing area is marked; receiving an operation of an operator on the splicing planning interface, and determining a region of interest in the candidate splicing region; and splicing the plurality of X-ray images based on the region of interest to obtain a spliced image of the spliced body position. According to the method, the region of interest is determined in the overlapping region or the candidate splicing region, and then the X-ray images are spliced based on the region of interest, so that the ghosting problem of anatomical structures / tissues concerned by an operator in the spliced image is reduced; and the definition of a ghosting problem of anatomical structures / tissues concerned by an operator in a spliced image is improved.

The invention provides a video time sequence consistency restoration method based on residual guided diffusion and cross-frame attention distillation, and relates to the technical field of video image restoration. The method comprises the following steps: constructing a degeneration-clear frame pair data set, and a teacher network and a student network containing a coding-decoding diffusion denoising backbone network; generating a noise-added image sequence; performing residual scale modulation through a residual scale prediction module to obtain a first restored feature sequence, and performing cross-frame attention fusion to obtain a second restored feature sequence; respectively outputting a teacher recovery result sequence and a student recovery result sequence; and constructing a total loss function containing reconstruction loss, distillation loss and time consistency loss in combination with the degradation-clear frame pair and the two types of restoration results, iteratively optimizing the student network, and restoring the video to be restored to obtain a target restored video sequence. According to the method, the ubiquitous problems that residual error utilization is not adaptive, time sequence modeling is insufficient and distillation lacks structure maintenance in an existing video restoration method are solved.

The single-mirror coaxial schlieren imaging method for observing vortex structure in a shock tube belongs to the technical field of fluid flow experiment test, and relates to a single-mirror coaxial schlieren imaging method for observing vortex structure in a shock tube. The method integrates a point light source and a blade to construct an integrated coaxial light source, and uses a schlieren mirror, a shock tube and a high-speed camera to build a single-mirror coaxial schlieren imaging system. In the system, the incident light of the point light source is reflected by the schlieren mirror, passes through the flow field in the shock tube and is focused on the tip of the blade. The blade is used to interfere with the focal point. Because the airflow density in the shock tube is different, the refractive index of the light is different, different schlieren gray images with different light and dark degrees are formed, and the high-speed camera is used to record the change of the vortex structure in the shock tube. The method is easy to debug and saves a lot of experimental debugging time. On the basis of ensuring the light intensity, the influence of ghosting and astigmatism is reduced, and the schlieren imaging of the vortex structure in the shock tube is accurate and clear.