30 results about "Image simplification" patented technology

A process is disclosed for displaying an image on a display screen and recognizing a user's input by the user's finger touch on a touch screen, the process including: a first step of displaying an original image; a second step of forming a screen pad, such that a reference image, which is a reduced duplicate of the original image, is displayed on the screen pad within the display screen, such that the reference image is overlaid with the original image, in a translucent mode that renders the original image visible through the reference image, or an opaque mode where the original image is not visible. The reference image is obtained by reducing the original image with or without image simplification. A third step, in response to the user touching the screen pad, detects finger position on the screen pad, and recognizes the user's input based on the detected finger position.

The disclosed positioning device for touch pad comprises: a display screen, a DC or image inductor above screen for a given distance and fir to shoot whole screen, a mono-color line light source set at any side of corner of screen whose beam closes to screen as much as possible and covers whole screen; and a high-bright light spot detection module connected between the DC or inductor and computer. This invention simplifies structure and algorithm, and improves anti- disturbance ability.

The invention provides a photovoltaic cell panel grid line defect detection method and system based on artificial intelligence, and relates to the field of artificial intelligence. The method comprises the following steps: acquiring an image of a single photovoltaic cell panel; constructing an image matrix according to prior color features in the photovoltaic cell panel image, and performing sliding window calculation on the image matrix to obtain a single-channel image; obtaining a compression value of each column of pixels of the single-channel image as a one-dimensional sequence representing the photovoltaic cell panel image; and inputting the one-dimensional sequence into a time convolution network to obtain the periodicity of the grid line in the photovoltaic cell panel image, and judging the defect position of the grid line according to the abnormal point. According to the invention, the semantic segmentation network type is adopted to identify the content of each column in the image, the image processing mode does not need to debug parameters, the periodicity of the content of the photovoltaic cell panel is utilized to adapt to cell panel images of various sizes, the preprocessing process is simplified, and the generalization ability is high. According to the invention, the influence of camera parameters and illumination intensity can be eliminated, automation is realized, and the detection efficiency and accuracy are improved.

The invention discloses a label construction method, system and device for a training model, and a storage medium, and the method comprises the steps: carrying out the label extraction of a figure image, simplifying a complex figure image into a key point two-dimensional coordinate or a two-dimensional mask, and carrying out the training of a generative adversarial neural network model (GAN model); different label images can be generated by simply modifying coordinate positions or two-dimensional mask shapes of key points, figure images corresponding to labels can be generated by inputting thetrained generative adversarial neural network model, and then videos are further synthesized, so that the operation process of figure video synthesis is greatly simplified, and the working efficiencyis improved; subsequently, a new label can be added as a limiting condition according to actual generation requirements, the label and the real image corresponding to the label are sent to the generative adversarial network model for training, and finally, the corresponding real image can be generated through expansion conditions. The method is widely applied to the technical field of image processing.

The invention discloses an access control system based on quantum password. The access control system comprises an image processing module, an information screening module and a face scanning model arranged on a door body, the output end of the face scanning module is electrically connected with the input end of an image miscellaneous point removing module, and the image miscellaneous point removing module is arranged in the image processing module; the output end of the image miscellaneous point removing module is electrically connected with the input end of an image simplification module, and the output end of the image simplification module is electrically connected with the first input end of an image matching module; and the image matching module is arranged in the information screening module. The access control system utilizes the face scanning module to scan the face of a user, then a mobile phone terminal of the user can receive the corresponding secret key, the user inputs the secret key received by the mobile phone terminal on a secret key input module, so that an electronic lock on the door body is opened, convenience is brought to the user, in addition, a quantum encryption technique is utilized for making the safety property be improved further.

The invention is suitable for the field of machine vision, and provides a calibration device and method and a depth camera, the device comprises a calibration plate, a moving part, a fixed support, a projection part, a camera and a processor; the calibration plate is provided with feature points and positioning information; the projection part is used for projecting the patterned light beam onto the calibration plate; the moving part is used for changing the position of the calibration plate; the camera and the projection component are fixed on the fixed bracket, and the camera is used for collecting the patterned light beam reflected by the calibration plate and generating a calibration image; and the processor is electrically connected with the moving part and the camera, and is used for receiving the calibration image acquired by the camera, controlling the moving part to move, and determining the internal parameters of the camera and the external parameters of the camera and the projection part based on the calibration image. The calibration images of the calibration plate at different positions are obtained by controlling the moving part, acquisition of the calibration material is simplified, and therefore the calibration process of relevant parameters of the camera is simplified.

The invention discloses a face data pair-oriented complete face reticulate pattern restoration method and system and a medium. The method comprises the following steps of combindinga face reticulate pattern image and a clean face image corresponding to the face reticulate pattern image into a picture in the horizontal dimension; taking a training set composed of the data pairs as input of a GAN network, and training to obtain a restoration model; and inputting a to-be-restored face reticulate pattern image into the restoration model to obtain a restored clean face image. The method is an end-to-end training mode, the step of reticulate pattern recognition is omitted, and the bottleneck of reticulate pattern segmentation-based restoration is broken through. The image with the reticulate pattern is used as noise data, the face image without the reticulate pattern can be directly generated through end-to-end adversarial training, the integrity of the restored face reticulate pattern imageis improved while the face reticulate pattern restoration process is simplified, and obvious restoration traces cannot appear.

The invention discloses a three-dimensional model attribute image simplification method and device based on rate distortion optimization. The method comprises the following steps: acquiring a three-dimensional model attribute image, equally dividing the whole attribute image into a plurality of image blocks, and determining a rate distortion curve of each image block; according to the rate distortion curve of each image block, under the limit of a given total number of pixels, using a Lagrange multiplier method to optimize and solve the down-sampling rate of each image block; and performing down-sampling on the corresponding image block by using bicubic interpolation according to the down-sampling rate of each image block to obtain a simplified attribute image. According to the method, the down-sampling rates of different areas of the attribute image are optimally determined, so that the simplified three-dimensional model with attributes is close to an original model as much as possible in appearance. The optimization process is not only based on the features of the 2D attribute image, but also considers the geometric features of the original and simplified 3D models and parameterization from the model surface to the attribute image. The attribute simplification method provided by the invention has universality.

In order to solve the defects in the prior art, the invention provides a mining area ecological restoration guidance system based on big data, which comprises an unmanned aerial vehicle patrol flight system and an analysis control device. The unmanned aerial vehicle patrol flight system comprises a patrol flight unmanned aerial vehicle and unmanned aerial vehicle platforms arranged at a plurality of preset positions in a mining area along a patrol flight route of the patrol flight unmanned aerial vehicle. The patrol flight unmanned aerial vehicle at least comprises a high-definition camera shooting towards the ground. The analysis control device comprises an unmanned aerial vehicle flight control module, an image analysis module and an ecological restoration guidance module. The unmanned aerial vehicle flight control module is used for controlling the patrol flight state of the patrol flight unmanned aerial vehicle, the image analysis module receives image data sent back by the patrol flight unmanned aerial vehicle, and a color gamut distribution diagram is formed after image simplification processing. And the ecological restoration guidance module is used for obtaining corresponding ecological restoration guidance information from the big database based on the color gamut distribution diagram and sending the information to the user. On one hand, the working efficiency is improved, and on the other hand, regular monitoring of ecological changes is realized.

The invention discloses a panoramic picture and video detection method and device and a readable storage medium. The panoramic picture detection method comprises the following steps: acquiring a panoramic picture, and presetting an angle parameter; performing angle comparison processing on the panoramic picture according to the angle parameter, and obtaining a parameter comparison result; performing image acquisition processing on the panoramic picture according to a parameter comparison result to obtain a plane picture; performing face detection processing on the plane picture to obtain a detection result, and outputting a face image according to the detection result; wherein the angle parameter at least comprises one of a roll angle parameter, a first overlook angle parameter and a first yaw angle parameter. And performing angle comparison processing on the panoramic picture, and performing image processing on the panoramic picture according to a parameter comparison result to obtain a plane picture. The plane picture is detected through the face detection processing to obtain the face image, the algorithm of face recognition of the panoramic picture is simplified, and the recognition accuracy is improved.

The embodiment of the invention provides a local image processing method and device, a vehicle-mounted system and a storage medium, and the method comprises the steps: displaying a plurality of local image options in a graphical interface; in response to a selection operation on the target local graph option, determining a to-be-displayed image type; generating a to-be-displayed image according to the type of the to-be-displayed image, and displaying the to-be-displayed image in the image interface; wherein the to-be-displayed image is an image displayed at a visual angle of a virtual camera, and a virtual camera coordinate system where the virtual camera is located is a coordinate system constructed by taking the position of the virtual camera under a vehicle body coordinate system as an origin of coordinates. According to the embodiment of the invention, the high-quality local image is displayed in the vehicle-mounted system, the local image can better meet the visual requirements of human eyes through the visual angle of the virtual camera, the local image is reversely mapped to the original image, and the mapping process is simplified.

The invention provides a panoramic picture face detection method. The panoramic picture face detection method comprises the following steps: S1, acquiring a panoramic picture, and presetting an angle parameter; s2, performing angle comparison processing on the panoramic picture according to the angle parameter, and obtaining a parameter comparison result; s3, performing image acquisition processing on the panoramic picture according to a parameter comparison result to obtain a plane picture; and S4, performing face detection processing on the plane picture to obtain a detection structure, and outputting a face image according to a detection result. According to the panoramic picture face detection method provided by the invention, angle comparison processing is performed on the panoramic picture, image processing is performed on the panoramic picture according to the parameter comparison result to obtain the plane picture, and face detection processing is performed on the plane picture to obtain the face image, so that a panoramic picture face recognition algorithm is simplified; and the recognition accuracy is improved.

According to the automatic rain removing method for the vehicle-mounted camera photographed image, raindrops are regarded as noise, firstly, a network model with input and output being images at different moments in the same rainy day scene is established, and rainy day noise distribution is obtained through the network model; and processing the rainy day image through a denoising network to obtain a rain-free effect picture. According to the invention, training is carried out on the basis that input and output are samples of rainy day images at different moments in the same scene, so that rainy day noise distribution is obtained, and rain removal processing is carried out on the rainy day images with the rainy day noise distribution. Thus, according to the invention, the training of the model does not need a clean rain-free image, the data processing difficulty is simplified, the feasibility is improved, and the accuracy degree of data processing is ensured.

The invention discloses a malicious PDF document intelligent detection method and system based on a graph structure. The method comprises the steps of obtaining a to-be-detected PDF document; analyzing a PDF document to obtain a graph structure feature of the document, performing classification importance sorting on all nodes of the graph structure feature, performing retention and merging processing according to a sorting result to obtain a simplified graph structure feature, and calculating a Laplacian matrix of the simplified graph structure feature as an input feature; inputting the input features into a pre-trained convolutional neural network model, and if the output is 1, determining that the document is a malicious document; if the output is 0, judging the document to be a benign document. The graph structure extraction and image simplification method has the advantages that the dimensionality of features is reduced, the training pressure of a deep learning model is relieved, and the efficiency of a system is improved; and according to the input features, the convolutional neural network model is utilized to detect and classify the documents or automatically train parameters of the model, so that the accuracy and usability of the system are improved.

The invention provides a PCA-based grey image simplification method. The method comprises the steps of extracting a grey image matrix of grey images; compressing the grey image matrix to obtain a primary compressed grey image matrix; compressing the primary compressed grey image matrix to obtain a secondary compressed grey image matrix; and restoring the secondary compressed grey image matrix to obtain simplified grey images. The invention furthermore provides an active interaction device, a non-transient readable storage medium and an apparatus, which are used for realizing the method. Redundant information of the grey images can be reduced; and spaces occupied in an electronic device by the grey images are reduced.