41results about How to "Image processing is simple" patented technology

The invention provides a face recognition method and a face recognition system. The method includes: Step S101: Segment the glasses in the glasses picture to obtain glasses templates; Step S102: Define face key points on the glasses-free face picture and define glasses on the glasses picture and the glasses template key points; and step S103: based on the face key points and the glasses key points, the glasses picture and the glasses template are adjusted to match the glasses-free face picture, and through the glasses The template superimposes the glasses in the glasses picture on the face picture without glasses to synthesize the face picture of people wearing glasses. The method can realize more rapid, simple and accurate recognition of the face of the person wearing glasses, so as to provide better recognition performance. The face recognition system also has the above-mentioned advantages.

A computational intensity correlation imaging autocollimator and measurement method, comprising: a light source, a digital micromirror array DMD, a controller, a beam splitter, a lens, a mirror of a test piece, a CCD and an intensity correlation calculation module; the synchronization of the controller The control module generates a synchronous clock signal to act on the digital micromirror array DMD and CCD. The intensity correlation calculation module simultaneously records the light field modulation matrix generated by the light field modulation module and the echo signal received by the CCD, and performs second-order intensity correlation calculation. The tiny rotation angle of the DUT is obtained through indirect calculation. The invention introduces the computational correlation imaging method in the quantum imaging technology into the design of the autocollimator, which can effectively reduce the measurement error caused by air disturbance and other factors, and improve the sensitivity and stability of the system.

The invention provides a multi-section dual-X-ray machine system. The multi-section dual-X-ray machine system comprises two columns of X-ray light sources, two columns of X-ray light surface receivers and a controller, wherein the X-ray light surface receivers of each column of X-ray light surface receivers are located on the same plane, one column of X-ray light sources is opposite to one column of X-ray light surface receivers, the two columns of X-ray light sources are crossed in orientation, and the crossed positions are detection zones. The X-ray light sources in each column of X-ray light sources respectively face the X-ray light surface receivers in the corresponding column of X-ray light surface receivers so that a receiving plate of one X-ray light surface receiver only receives X-ray light of one corresponding X-ray light source. The controller controls all the X-ray light sources to transmit primary X-ray light simultaneously or in a time-share mode to obtain surface images of corresponding sections of corresponding light surfaces on the X-ray light surface receivers, and accordingly multi-section dual-light-surface pre-detected object surface images are obtained. The radiation dosage is small, whole-body images of a human body at a standing position in two azimuths can be obtained, medical analysis and planning are promoted, and the scanning speed is very high.

The invention discloses an image processing method and an electronic device. The method includes acquiring at least three images output through at least three image acquisition units respectively; synchronously controlling the acquired at least three images to obtain the synchronized control result; performing the distortion correction on the at least three images to to obtain at least three correction images based on the synchronized control result; and splicing the at least three correction images to obtain a target image.

A DLS model-based high-definition processing method and system for cadre personnel files are disclosed, an image processing DLS model is constructed to record and analyze the image boundary information. The square of L2 norm of the gradient is used as a regular constraint condition to constrain the gradient of the boundary. The regularization denoising algorithm and the image smoothing distribution logic are introduced to minimize the total variation while obtaining the best image. Based on the image processing DLS model, the cadre personnel files are scanned to obtain scanned images and stored, and meanwhile, the scanned images are transmitted to the server for high-definition processing to obtain high-definition images. The invention innovates the image high-definition processing technology, and avoids the problems of image color, shape distortion, discrimination reduction and the like caused by the unreasonable application of the image acquisition technology.