14 results about "Gray level image" patented technology

The application discloses a kind of based on branch current's latch detection method, device and latch protection circuit.The method includes obtaining the branch current from a protection branch;Branch current is converted into time-frequency domain signal;Time-frequency domain signal is structured as gray scale image;Extract the texture feature of gray scale image, whether the corresponding branch occurs latch and micro-latch is judged using pre-trained machine learning model.The method of this embodiment, directly for branch current detection, detection point quantity is exponentially decreased, improve the utilization of PCB board, also reduce the single board power consumption.In addition, the branch current of the application is structured as gray scale image, extracts the current texture feature in image, and is judged by pre-set machine learning model, the accuracy of identifying latch and micro-latch is high, supports the detection of various types of devices, so as to effectively protect the single board circuit in orbit operation, prevent it from being damaged due to latch or micro-latch.

The present application relates to the technical field of image enhancement, in particular to a contrast enhancement method for nylon thread surface microscopic image, comprising the following steps: obtaining a nylon thread surface microscopic gray image, calculating the horizontal and vertical gradient values of the pixel points of the nylon thread surface microscopic gray image, constructing a pixel-level structure tensor matrix, performing eigenvalue decomposition operation on the pixel-level structure tensor matrix, and generating a nylon thread texture coherence distribution factor. In the present application, the nylon thread roughness gain mapping points are multiplied by the detail component and reconstructed with the background component to output a high-contrast enhanced image of the nylon thread surface, so that the contrast enhancement is concentrated on the real texture details and the risk of pseudo-texture caused by background gray drift and excessive enhancement is reduced, while the texture direction consistency enhancement, local roughness difference enhancement, and visual readability improvement after the separation of details and background are realized.

This application relates to a method for enhancing the quality of wafer dark-field defect detection images. The method includes: under conditions of wafer micro-vibration, separating the dark-field scattered light from the wafer surface into two orthogonal polarization components, an s-polarization channel and a p-polarization channel, using a polarization beam splitter; acquiring time-series image sequences corresponding to the s-polarization channel and the p-polarization channel respectively; processing the time-series image sequences to obtain a standard image sequence; calculating the time-series gray-level variation coefficient of each pixel position in the s-polarization channel and the p-polarization channel respectively, and calculating the time-series gray-level cross-correlation coefficient F between the s-polarization channel and the p-polarization channel; determining the background roughness scattering region and the defect scattering region based on the time-series gray-level variation coefficient, and forming suppression weights based on the determination results; acquiring a residual gray-level image; and performing nonlinear gray-level stretching on the residual gray-level image to obtain a defect detection image with enhanced signal-to-noise ratio. This invention can improve the image signal-to-noise ratio.

This invention discloses a ground-penetrating radar (GPR) image recognition method based on adaptive threshold segmentation, relating to the fields of geophysical exploration and digital image processing. The method includes: S1, acquiring raw image data obtained from GPR detection and preprocessing it to enhance target signals and suppress noise; S2, generating an individual optimal threshold for each preprocessed GPR grayscale image, resulting in an optimal threshold set; S3, performing mean processing on this set to calculate a globally unified threshold that reflects the segmentation level of the entire detection area; S4, using the globally unified threshold to binarize all raw input grayscale images, and then performing morphological opening operations on the binarized images; S5, performing connected component analysis on the opened images and applying area thresholding again to filter out connected components smaller than a second preset area threshold, outputting the final segmentation result image; This method features fast and accurate recognition.

The application discloses a digital heat manipulation method based on image programming and relates to the field of computer-aided (CAX) technology, characterized in that for a heat manipulation process involving multiple heat sources, in order to realize accurate control of a target surface temperature field, the process is first discretized according to a given time step to obtain the target temperature field in each time step; the temperature field of the target surface is characterized by a gray-scale image, the image is processed in a fuzzy manner in combination with a heat transfer equation to simulate the evolution process of the temperature field over time, and the required theoretical optimal heating mode is reversely solved; and the input power corresponding to each heat source is calculated by a characteristic decomposition algorithm to approximate the heating mode. Thus, the input power state corresponding to each heat source at each moment is determined, and digital heat manipulation is realized. The application provides a new heat manipulation idea, provides more efficient, accurate and unified operation, and lays a theoretical and technical foundation for automatic and accurate control of a microwave solidification manufacturing process.

The application relates to the technical field of data processing, in particular to a non-contact polyurethane foam plate detection method, which comprises the following steps: data processing is performed on each pixel point in a Gaussian filter window to obtain a gray level co-occurrence matrix; a Gaussian filter variance value is obtained according to the gray level co-occurrence matrix; autocorrelation of each neighborhood pixel point in the Gaussian filter window corresponding to each pixel point is obtained; weight values of each neighborhood pixel point are obtained according to coordinate values and autocorrelation of each neighborhood pixel point; initial Gaussian values of each neighborhood pixel point are obtained according to the Gaussian filter variance value; target Gaussian values are obtained according to the weight values and the initial Gaussian values; a target gray level image after Gaussian filter smoothing is obtained according to the target Gaussian values and gray level values of each neighborhood pixel point in the Gaussian filter window corresponding to each pixel point; and a scratch defect edge is obtained according to the target gray level image after Gaussian filter smoothing. The application can improve the precision of polyurethane foam plate abnormality detection.

The present application relates to the technical field of industrial vision detection, in particular to a myopia-preventing paper quality detection system based on industrial vision, comprising: a multi-angle image acquisition module for acquiring a paper surface gray image under multi-angle illumination environment; a frequency domain analysis module for analyzing frequency distribution characteristics and calculating local spectrum entropy to quantify diffuse reflection randomness; a texture analysis module for constructing a gray level co-occurrence matrix to extract texture energy difference degree and represent light reflection isotropic characteristics; a glossiness analysis module for calculating glossiness consistency index to evaluate light softness degree; and a defect identification module for fusing local spectrum entropy, texture energy difference degree and glossiness consistency index to construct a multi-dimensional texture feature set, identifying functional diffuse reflection abnormal areas and generating a detection result; the present application quantifies visual comfort from a microscopic optical perspective, ensuring functional quality of light softness of myopia-preventing paper.

The application discloses a network intrusion detection method and system based on deep learning, relates to the technical field of network security, and comprises the following steps: collecting network flow data by using a network packet capturing tool, constructing a space-time matrix by using a mapping matrix construction method, converting the space-time matrix into a gray image, dynamically adjusting by using a dynamic weight adjustment formula, constructing a topological complex, screening important topological features, respectively performing enhancement processing on the important topological features, fusing by using a weighted average method, and generating a fused gray image; segmenting by using a space-time mapping generation method, calculating a DTW (Dynamic Time Warping) distance by using a dynamic time warping calculation, marking a mutation point segment, calculating the evolution fitness of the mutation point segment by using an evolution fitness formula, and identifying an intrusion behavior. The neural biological signal processing technology is combined with the topological data analysis, the processability of data and the identifiable property of features are improved, the evolution fitness formula is used for analyzing the mutation point segment, and the detection precision and the adaptability of the intrusion behavior are enhanced.

The application relates to the technical field of chip defect identification, and discloses a wafer chip defect detection method, device, equipment and medium. In the method, a Mamba space modeling network is used, so that when a high-resolution original gray image of a wafer chip is processed, the calculation complexity is reduced from a square level of an existing method to a linear level, the inference speed is improved, and the demand of real-time detection of an industrial production line is met. The spatial propagation mechanism of the Mamba space modeling network effectively captures the spatial propagation characteristics of wafer chip defects, improves the identification accuracy and spatial consistency of systematic defects and continuous defects, the learnable spline base function of the KAN network enhances the interpretability of the feature transformation process, the decision basis of the detection result is transparent and traceable, and quality auditing and process improvement are facilitated.

Embodiments of the present specification provide a point cloud data-based image recognition method, device and equipment and a storage medium, wherein the method comprises: obtaining point cloud data of a to-be-recognized object; dividing the point cloud data to obtain a plurality of matrices, and calculating an average value of distances between all elements in each matrix as a value of the matrix; generating a gray-scale image of the to-be-recognized object according to the value of each matrix; and inputting the gray-scale image of the to-be-recognized object into a convolutional neural network model to recognize the to-be-recognized object. Embodiments of the present specification can determine the specific morphology of the to-be-recognized object, and improve the accuracy of recognition.

The application discloses a target recognition method based on a neuromorphic visual sensor, and belongs to the technical field of computer vision and artificial intelligence, and comprises the steps of collecting event data and a gray image; adopting a variable event window strategy and a time surface mechanism to construct an event frame for the event data; inputting the event frame and the gray image after fusion into a VGG11 network to extract multi-level joint features; matching real-time joint features with template features in a pre-constructed target feature template library, and realizing target classification and positioning according to a similarity score map; and the application adopts the above target recognition method based on the neuromorphic visual sensor, fuses the high time resolution of the event data and the spatial structure information of the gray image, and significantly improves the target recognition precision in a complex dynamic scene.

The application relates to the technical field of robot object picking methods, and particularly discloses an object picking method based on correlation analysis and robot operation, which comprises the following steps: step one, obtaining an ROI image through an image acquisition device; step two, performing image processing and target detection on the ROI image, specifically, importing the ROI image into software to process the ROI image into a gray scale image, then detecting an object image of an object in the ROI gray scale image and positioning the object image; step three, calculating the center coordinates of the object image center in the ROI gray scale image, and converting the center coordinates into actual distance coordinates of a spatial actual distance; and step four, using the actual distance coordinates in a robot control instruction, and the robot performing actions according to the robot control instruction. The method can greatly improve the automatic adaptability of the robot by optimizing the computer vision technology algorithm for the robot arm to pick and place objects based on correlation analysis, and can simplify the calculation and processing process and improve the accuracy.