Edge detection method and device based on characteristics of self-sensitized neuron material
The self-synaptic neuronal material-based edge detection method addresses the limitations of existing methods by enhancing adaptability and reducing computational costs, enabling efficient edge detection across various scenarios.
Patent Information
- Application Number
- CN202410051339.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-12
- Publication Date
- 2025-07-15
AI Technical Summary
The existing edge detection method has a single application scenario, many adjustment parameters, high calculation cost, and it is difficult to efficiently obtain good edge detection results.
By reading the original image data of the target, decompose the RGB channel image data, calculate the gradient information, and input the gradient information into the self-sensitized neuron within the preset simulation time, adjust the neuron state, calculate the pulse excitation frequency, and convert it into grayscale pixel values for edge detection.
It breaks through the limitations of traditional operator application scenarios, improves the environmental adaptability of edge recognition technology, reduces the number of parameters, improves the computing efficiency, and makes edge recognition simpler and faster.
Smart Images
Figure CN120318261A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of edge detection technology, and particularly to an edge detection method and device based on the characteristics of self-sensitized neuron materials. Background Art
[0002] Edge detection, as a key task in the fields of computer vision and image processing, is often used to identify object boundaries or features in images.
[0003] Existing edge detection methods can use different operators such as Sobel and Roberts to provide different methods for extracting edges and features; however, although using these operators can detect strong edges in images, they usually lose weak details, are more sensitive to noise, or may over-segment the images, with a single application scenario and great limitations in use.
[0004] In addition, existing edge detection methods such as multi-stage algorithms like the Canny algorithm can effectively reduce the influence of noise and generate detailed edges through steps such as Gaussian smoothing, calculating gradients, non-maximum suppression, and edge tracking; however, this algorithm is relatively complex, requires adjusting multiple parameters to obtain the best results, has a high computational cost, and is not suitable for real-time applications.
[0005] In summary, existing edge detection methods using operators alone cannot obtain good edge detection effects, have a single application scenario, are difficult to be applicable to different scenarios simultaneously, and in addition, some edge detection methods require many parameter adjustments and have a high computational cost, which urgently need to be solved. Summary of the Invention
[0006] This application provides an edge detection method and device based on the characteristics of self-sensitized neuron materials to solve the problems of existing edge detection methods having a single application scenario, many parameter adjustments, high computational costs, and being difficult to efficiently obtain good edge detection effects.
[0007] The first aspect of the embodiments of this application provides an edge detection method based on the characteristics of self-sensitized neuron materials, including the following steps: reading the original image data of the target, and decomposing the RGB channel image data in the original image data; calculating the gradient information of the RGB channel image data, and based on the gradient information of the RGB channel image data, obtaining the gradient information of the original image data, and within a preset simulation time, repeatedly inputting the gradient information of the original image data into a preset self-sensitized neuron, adjusting the state of the self-sensitized neuron, and calculating the pulse excitation frequency of the self-sensitized neuron, and converting the pulse excitation frequency into a gray pixel value to perform visualization of the preset target edge detection according to the gray pixel value.
[0008] Optionally, in an embodiment of the present application, calculating the gradient information of the RGB channel image data and obtaining the gradient information of the original image data based on the gradient information of the RGB channel image data includes: performing a Gaussian smoothing operation on the RGB channel image data to obtain an RGB channel Gaussian blurred image; calculating the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image; adding the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image to obtain the gradient information of the original image data of the preset target.
[0009] Optionally, in an embodiment of the present application, before adjusting the state of the self-sensitizing neuron and calculating the pulse firing frequency of the self-sensitizing neuron, it further includes: preprocessing the gradient information of the original image data to obtain the input data of the self-sensitizing neuron; within the preset simulation time, repeatedly inputting the input data into the self-sensitizing neuron, and updating the membrane potential of the self-sensitizing neuron based on a preset self-sensitizing neuron potential calculation formula to obtain the latest neuron potential; inputting the latest neuron potential into a preset neuron firing function to determine whether the self-sensitizing neuron fires.
[0010] Optionally, in an embodiment of the present application, adjusting the state of the self-sensitizing neuron includes: when the self-sensitizing neuron fires, obtaining the historical firing pulse number of the self-sensitizing neuron, updating the preset simulation time to obtain a new simulation time, and determining whether the new simulation time is a preset checkpoint; if the new simulation time is the preset checkpoint, determining whether the preset checkpoint and the historical firing record meet a preset change condition; if the preset change condition is met, switching the state of the self-sensitizing neuron to a preset state.
[0011] Optionally, in an embodiment of the present application, calculating the pulse firing frequency of the self-sensitizing neuron includes: recording the pulse number of the self-sensitizing neuron; calculating the pulse firing frequency of the self-sensitizing neuron according to the pulse number and the preset simulation time.
[0012] The second aspect of the present application provides an edge detection device based on the characteristics of self-sensitized neuron materials, including: a decomposition module for reading the original image data of the target and decomposing the RGB channel image data in the original image data; a calculation module for calculating the gradient information of the RGB channel image data and obtaining the gradient information of the original image data based on the gradient information of the RGB channel image data; and a visualization module for repeatedly inputting the gradient information of the original image data into a preset self-sensitized neuron within a preset simulation time, adjusting the state of the self-sensitized neuron, calculating the pulse excitation frequency of the self-sensitized neuron, and converting the pulse excitation frequency into a grayscale pixel value to perform visualization of the preset target edge detection according to the grayscale pixel value.
[0013] Optionally, in an embodiment of the present application, the calculation module includes: a smoothing unit for performing Gaussian smoothing operation on the RGB channel image data to obtain an RGB channel Gaussian blurred image; a first operation unit for calculating the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image; and a second operation unit for summing the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image to obtain the gradient information of the original image data of the preset target.
[0014] Optionally, in an embodiment of the present application, it further includes: a preprocessing module for preprocessing the gradient information of the original image data before adjusting the state of the self-sensitized neuron and calculating the pulse excitation frequency of the self-sensitized neuron to obtain the input data of the self-sensitized neuron; an update module for repeatedly inputting the input data into the self-sensitized neuron within the preset simulation time and updating the membrane potential of the self-sensitized neuron based on a preset self-sensitized neuron potential calculation formula to obtain the latest neuron potential; and a determination module for inputting the latest neuron potential into a preset neuron excitation function to determine whether the self-sensitized neuron is excited.
[0015] Optionally, in an embodiment of the present application, the visualization module includes: a judgment unit for obtaining the historical excitation pulse number of the self-sensitized neuron when the self-sensitized neuron is excited, updating the preset simulation time to obtain a new simulation time, and judging whether the new simulation time is a preset checkpoint; an analysis unit for determining whether the preset checkpoint and the historical excitation record meet a preset change condition if the new simulation time is the preset checkpoint; and a switching unit for switching the state of the self-sensitized neuron to a preset state if the preset change condition is met.
[0016] Optionally, in an embodiment of the present application, the visualization module further includes: a recording unit for recording the number of pulses of the self-sensitizing neuron; a third operation unit for calculating the pulse excitation frequency of the self-sensitizing neuron according to the number of pulses and the preset simulation time.
[0017] An embodiment of the third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the edge detection method based on the material characteristics of the self-sensitizing neuron as described in the above embodiments.
[0018] An embodiment of the fourth aspect of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the program is executed by a processor, it implements the above edge detection method based on the material characteristics of the self-sensitizing neuron.
[0019] Therefore, the embodiments of the present application have the following beneficial effects:
[0020] The embodiments of the present application can read the original image data of the target, decompose the RGB channel image data in the original image data; calculate the gradient information of the RGB channel image data, and based on the gradient information of the RGB channel image data, obtain the gradient information of the original image data; within the preset simulation time, repeatedly input the gradient information of the original image data into the preset self-sensitizing neuron, adjust the state of the self-sensitizing neuron, and calculate the pulse excitation frequency of the self-sensitizing neuron, and convert the pulse excitation frequency into a grayscale pixel value, so as to perform the visualization of the preset target edge detection according to the grayscale pixel value, thereby breaking through the limitations of the application scenarios of traditional operators, improving the environmental adaptability of the edge recognition technology, reducing the number of parameters, improving the calculation efficiency, and making the edge recognition simpler, more convenient and faster. Thus, the problems of the existing edge detection methods such as single application scenario, many adjustable parameters, high calculation cost, and difficulty in efficiently obtaining good edge detection effects are solved.
[0021] The additional aspects and advantages of the present application will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present application. Description of the Drawings
[0022] The above and / or additional aspects and advantages of the present application will become obvious and easy to understand from the following description of the embodiments in conjunction with the drawings, where:
[0023] Figure 1 It is a flowchart of an edge detection method based on the material characteristics of the self-sensitizing neuron according to an embodiment of the present application;
[0024] Figure 2Execution logic schematic diagram of an edge detection method based on the characteristics of self-sensitized neuron materials provided by an embodiment of the present application;
[0025] Figure 3 Schematic diagram for comparing the effects of different edge detection methods after changing from a bright environment to a dark environment provided by an embodiment of the present application;
[0026] Figure 4 Schematic diagram for comparing the effects of different edge detection methods in a gradually changing environment provided by an embodiment of the present application;
[0027] Figure 5 Schematic diagram for comparing the car edge detection results in environments with different light intensities provided by an embodiment of the present application;
[0028] Figure 6 Schematic diagram for comparing the structural features of self-sensitized edge neurons and traditional edge recognition methods when entering a dark environment provided by an embodiment of the present application;
[0029] Figure 7 Example diagram of an edge detection device based on the characteristics of self-sensitized neuron materials according to an embodiment of the present application;
[0030] Figure 8 Schematic structural diagram of an electronic device provided by an embodiment of the present application.
[0031] Among them, 10 - edge detection device based on the characteristics of self-sensitized neuron materials, 100 - decomposition module, 200 - calculation module, 300 - visualization module, 801 - memory, 802 - processor, 803 - communication interface. Detailed implementation manners
[0032] The embodiments of the present application are described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present application, and should not be construed as a limitation to the present application.
[0033] The edge detection method and device based on the characteristics of self-sensitized neuron materials in the embodiments of the present application will be described below with reference to the accompanying drawings. In view of the problems mentioned in the above background art, the present application provides an edge detection method based on the characteristics of self-sensitized neuron materials. In this method, by reading the original image data of the target, the RGB channel image data in the original image data is decomposed; the gradient information of the RGB channel image data is calculated, and based on the gradient information of the RGB channel image data, the gradient information of the original image data is obtained; within a preset simulation time, the gradient information of the original image data is repeatedly input into a preset self-sensitized neuron to adjust the state of the self-sensitized neuron, and the pulse excitation frequency of the self-sensitized neuron is calculated. The pulse excitation frequency is converted into a grayscale pixel value to perform visualization of preset target edge detection according to the grayscale pixel value, thereby breaking through the limitations of the application scenarios of traditional operators, improving the environmental adaptability of edge recognition technology, reducing the number of parameters, improving the calculation efficiency, and making edge recognition simpler, more convenient and faster. Thus, the problems of the existing edge detection methods, such as single application scenario, many adjustable parameters, high calculation cost, and difficulty in efficiently obtaining good edge detection effects, are solved.
[0034] Specifically, Figure 1 FIG. is a flowchart of an edge detection method based on the characteristics of self-sensitized neuron materials provided by an embodiment of the present application.
[0035] As Figure 1 shown, the edge detection method based on the characteristics of self-sensitized neuron materials includes the following steps:
[0036] In step S101, the original image data of the target is read, and the RGB channel image data in the original image data is decomposed.
[0037] In the embodiments of the present application, the original image of the target can be read first, and the original image is decomposed, as Figure 2 shown, to obtain monochromatic picture data of three RGB channels, and the obtained image data is saved, so as to provide a reliable data basis for subsequent image gradient calculation and neuron processing.
[0038] In step S102, the gradient information of the RGB channel image data is calculated, and based on the gradient information of the RGB channel image data, the gradient information of the original image data is obtained.
[0039] After obtaining the RGB channel image data in the original image data, further, embodiments of the present application can also calculate the gradient information of the RGB channel image data through the Sobel operator, so as to obtain gradient information such as the gradient magnitude and gradient direction of each pixel in the original image. Among them, the gradient magnitude represents the intensity of the pixel value change, and the gradient direction is the direction of the fastest change, thus effectively ensuring the performance of the subsequent edge detection of the self-sensitizing neuron material characteristics.
[0040] Optionally, in an embodiment of the present application, the gradient information of the RGB channel image data is calculated, and based on the gradient information of the RGB channel image data, the gradient information of the original image data is obtained, including: performing a Gaussian smoothing operation on the RGB channel image data to obtain an RGB channel Gaussian blurred image; calculating the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image; adding the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image to obtain the gradient information of the original image data of a preset target.
[0041] It should be noted that the specific steps for embodiments of the present application to obtain the gradient information of the original image data are as follows:
[0042] (1) Perform a Gaussian smoothing operation on the RGB channel image:
[0043] Convolve each pixel in the RGB channel image with a Gaussian kernel, and the resulting pixel value is the weighted average of the surrounding pixel values fused with Gaussian weights, thereby reducing noise while blurring small changes in the image;
[0044] (2) Calculate the gradient magnitude and direction of the RGB channel image:
[0045] Embodiments of the present application can calculate the image gradient information through the Sobel operator. Among them, the Sobel operator can be a convolution kernel used to detect the edge direction in an image. For horizontal and vertical edges, the Sobel operator has two convolution kernels respectively. By applying these convolution kernels to the image, horizontal and vertical gradient images can be obtained;
[0046] (3) Calculate the gradient information of the original image:
[0047] Add the gradient information of the image data of the three RGB channels to obtain the gradient information of the entire original image.
[0048] It should be noted that in the actual execution process, in addition to using the Sobel operator to calculate the image gradient information, those skilled in the art can also use other operators such as the Canny operator to calculate the image gradient information, which does not affect the superiority of the subsequent self-sensitizing neuron in edge detection applications, and is not specifically limited herein.
[0049] Accordingly, the embodiments of the present application calculate the image gradient information by using Gaussian smoothing and Sobel operators, thereby effectively reducing image noise, improving image quality, and providing reliable data support for the subsequent processing of self-sensitizing neurons.
[0050] In step S103, within a preset simulation time, the gradient information of the original image data is repeatedly input into a preset self-sensitizing neuron, the state of the self-sensitizing neuron is adjusted, and the pulse firing frequency of the self-sensitizing neuron is calculated. The pulse firing frequency is converted into a gray pixel value for visualizing a preset target edge detection according to the gray pixel value.
[0051] After obtaining the gradient information of the original image data, further, the embodiments of the present application can also input the gradient information of the original image data into a neuron layer, and the self-sensitizing neuron processes the gradient information of the original image data, thereby realizing the visualization of a preset target edge detection.
[0052] Optionally, in an embodiment of the present application, before adjusting the state of the self-sensitizing neuron and calculating the pulse firing frequency of the self-sensitizing neuron, it further includes: preprocessing the gradient information of the original image data to obtain the input data of the self-sensitizing neuron; within a preset simulation time, repeatedly inputting the input data into the self-sensitizing neuron, and updating the membrane potential of the self-sensitizing neuron based on a preset self-sensitizing neuron potential calculation formula to obtain the latest neuron potential; inputting the latest neuron potential into a preset neuron firing function to determine whether the self-sensitizing neuron fires.
[0053] It should be noted that after obtaining the gradient information of the original image data, the embodiments of the present application can perform preprocessing operations on the gradient information, and the specific process is as follows:
[0054] 1. Convert the data type to Tensor type to ensure compatibility with the PyTorch framework;
[0055] 2. Standardize the data to ensure that it falls within an appropriate range during the processing of the self-sensitizing neuron:
[0056] (1) Subtract the minimum value 0 from all elements in the Tensor, and at the same time use the torch.max() function to obtain the maximum value in the Tensor elements, and divide by the maximum value to normalize all elements to the interval [0,1];
[0057] (2) Multiply by a suitable amplification factor to ensure matching with the neuron firing probability function during the calculation process.
[0058] Furthermore, embodiments of the present application can set variables to save the two-dimensional information (width and height) of the preprocessed input data, and perform dimensionality reduction on it to obtain one-dimensional data, which is then input into the neuron layer to simulate the forward propagation process in the spiking neural network.
[0059] During the current simulation time, embodiments of the present application can repeatedly input the preprocessed input data into the neuron layer. In 100 simulation calculation steps, embodiments of the present application can cyclically input the data into the neuron layer. In a single time (or called a simulation time step), the neuron can regard the input data as the stimulation intensity, calculate the latest neuron potential under the action of the corresponding intensity, and calculate whether the neuron fires according to the firing function of the neuron in the neuron firing function. The specific process is described as follows:
[0060] 1. At the current simulation time step, update the neuron potential corresponding to the input stimulus. The calculation formula for the neuron potential is:
[0061]
[0062] In the formula, represents the change rate of the membrane potential with time, τ in represents the input time constant, τ lk represents the leakage time constant, V inject represents the input voltage; among them, τ in , τ lk are determined by the physical properties of the neuron material and can be obtained through specific tests. Their specific calculation formulas are as follows:
[0063]
[0064]
[0065] Among them, C m is the capacitance value, representing the capacitance of the circuit or system. The capacitor is an element that stores charge and plays an important role in the transmission of input signals and the response speed of the system; R d , R s are the resistance values, representing the series resistance and the parallel resistance respectively. The resistor is an element that controls the flow of current and affects the damping and leakage characteristics of the system.
[0066] Subsequently, an embodiment of the present application can use the latest neuron potential as an input, calculate the firing probability of the neuron at the current simulation time step according to the firing probability function of the neuron state corresponding to the current latest neuron potential, and use the firing probability as an input. Through the torch.bernoulli() function, it is obtained whether the latest neuron fires. An output of 1 indicates that the corresponding neuron fires, and an output of 0 indicates that the corresponding neuron does not fire. Thus, the calculation of the stimulus intensity by the neuron and the update of the membrane potential can be realized, and it can be accurately determined whether the latest neuron fires, providing guidance and basis for the subsequent state adjustment of the self-sensitized neuron.
[0067] Optionally, in an embodiment of the present application, adjusting the state of the self-sensitized neuron includes: when the self-sensitized neuron fires, obtaining the historical firing pulse count of the self-sensitized neuron, updating the preset simulation time to obtain a new simulation time, and determining whether the new simulation time is a preset checkpoint; if the new simulation time is a preset checkpoint, determining whether the preset checkpoint and the historical firing record meet the preset change condition; if the preset change condition is met, switching the state of the self-sensitized neuron to the preset state.
[0068] After the corresponding calculations of all corresponding neurons in the Tensor are completed, the calculation results are output, and the pulse firing results of all neurons are accumulated to obtain the total number of firing pulses up to the current simulation time step.
[0069] Those skilled in the art should understand that in the process of calculating pulses, it is more important to complete the state adjustment of the self-sensitized neuron. The self-sensitized neuron can reduce the firing difficulty of the neuron when continuously receiving stimuli without firing pulses. In the self-sensitized neuron, different neuron states correspond to different neuron firing difficulties, that is, different neuron "thresholds".
[0070] Specifically, in the simulation time step calculation loop, after the accumulation of firing pulses is completed, an embodiment of the present application can judge the current simulation time step according to the updated simulation time step information to determine whether the current simulation time is a preset checkpoint.
[0071] The checkpoints in the embodiments of the present application can be set to the 2nd, 3rd, 6th, 12th, 18th, and 50th simulation time steps. If the checkpoint requirements are met, it is determined according to the current neuron historical firing record. The specific judgment process is as follows:
[0072] If the historical firing pulse count of the neuron is 0, change the state of the neuron corresponding to the neuron layer.
[0073] Among them, when the checkpoint is 1 and the simulation time step is 2, the change condition is met, and the neuron state is switched to state 1;
[0074] When the checkpoint is 2 and the simulation time step is 3, if the change condition is satisfied, the neuron state switches to state 2;
[0075] When the checkpoint is 3 and the simulation time step is 6, if the change condition is satisfied, the neuron state switches to state 3;
[0076] When the checkpoint is 4 and the simulation time step is 12, if the change condition is satisfied, the neuron state switches to state 4;
[0077] When the checkpoint is 5 and the simulation time step is 18, if the change condition is satisfied, the neuron state switches to state 5;
[0078] When the checkpoint is 6 and the simulation time step is 50, if the change condition is satisfied, the neuron state switches to the ground state.
[0079] It should be noted that the above data and the firing functions corresponding to different neuron states are all obtained from the device self - sensitization characteristic test. Among them, the firing functions of the neuron ground state and states 1 - 5 are as follows:
[0080] Ground state:
[0081] State 1:
[0082] State 2:
[0083] State 3:
[0084] State 4:
[0085] State 5:
[0086] Thus, the embodiments of the present application can effectively avoid the interference of noise data by switching the neuron states at different checkpoints and resetting the ground state of the neuron state at a certain checkpoint.
[0087] Optionally, in an embodiment of the present application, calculating the pulse firing frequency of the self - sensitizing neuron includes: recording the number of pulses of the self - sensitizing neuron; calculating the pulse firing frequency of the self - sensitizing neuron according to the number of pulses and the preset simulation time.
[0088] In the actual execution process, the pulse calculation process is accompanied by the sending of pulses. After each calculation, the number of pulses of the neuron can be recorded and accumulated. After the end of the simulation time loop, the embodiments of the present application can calculate the pulse firing frequency of the neuron, that is, the total number of fired pulses of the neuron within the simulation time loop divided by the simulation time step. Different firing intensities characterize the responses of neurons to different input stimuli.
[0089] After completing the calculation of the neuron layer, the embodiments of the present application can reshape the data according to the previously saved two-dimensional information and convert it into the uint8 type in the numpy library as the output of the algorithm. Finally, according to the gray mapping strategy, the pulse firing frequency is converted into gray pixel values, so as to obtain a reliable edge detection result and realize the visualization of the edge detection result. Moreover, the parameters are less adjusted and the calculation cost is lower, which can be applied to different application scenarios.
[0090] The following will analyze the performance of the edge detection method based on the characteristics of self-sensitizing neuron materials of the present application in conjunction with the accompanying drawings.
[0091] Figure 3 is a schematic diagram of the effect comparison of different edge detection methods after changing from a bright environment to a dark environment. As Figure 3 shown, by using traditional and the edge detection method of the present application to identify the detailed edge information of the license plate of a vehicle, when the two algorithms respectively identify from a bright environment picture to a dark environment picture, it can be seen that the traditional method cannot adapt to this environmental change, resulting in a sharp decline in the effect, while the self-sensitizing neuron edge recognition method can autonomously adapt to different environmental conditions and break through the limitation of the single application scenario of the traditional method.
[0092] Figure 4 is a schematic diagram of the effect comparison of different edge detection methods in a gradually changing environment. As Figure 4 shown, in a gradually changing environment, compared with the traditional edge detection method, the superiority of the edge detection method of the present application is more intuitive and obvious.
[0093] Figure 5 is a schematic diagram of the comparison of car edge detection results in different light intensity environments. As Figure 5 shown, compared with traditional edge recognition, the self-sensitizing neuron edge recognition has similar edge detection results and shows similar detection performance in a strong light environment. However, in a dark environment: under weak light intensity, the detection results output by the self-sensitizing neuron have a great improvement compared with traditional edge detection, and there is more complete object structure information, and its structure information is basically stable compared with the detection in a strong light environment, fully demonstrating its excellent performance level.
[0094] In addition, Figure 6Schematic diagram for comparing the structural features between self-sensitized edge neurons and traditional edge recognition methods when entering a dark environment. As Figure 6 shown, it shows that when more edge detections are in a dark environment, self-sensitized edge neurons can represent more complete structural features compared to traditional edge recognition methods.
[0095] It can be understood that the embodiments of the present application can effectively solve the limitation that traditional methods cannot be compatible with the simplicity and self-adaptability of algorithms by using self-sensitized neurons and applying them to edge detection scenarios in combination with biological intelligence characteristics, break through the limitations of the application scenarios of traditional operators. At the same time, due to the algorithm self-adaptability inspired by the self-sensitization ability of neurons, edge recognition is simple, convenient, fast, with fewer parameters to adjust and lower computational costs, and can flexibly adapt to various changing environments.
[0096] According to the edge detection method based on the material characteristics of self-sensitized neurons proposed by the embodiments of the present application, by reading the original image data of the target, the RGB channel image data in the original image data is decomposed; the gradient information of the RGB channel image data is calculated, and based on the gradient information of the RGB channel image data, the gradient information of the original image data is obtained; within a preset simulation time, the gradient information of the original image data is repeatedly input into a preset self-sensitized neuron, the state of the self-sensitized neuron is adjusted, and the pulse firing frequency of the self-sensitized neuron is calculated. The pulse firing frequency is converted into a grayscale pixel value to perform visualization of preset target edge detection according to the grayscale pixel value, thereby breaking through the limitations of the application scenarios of traditional operators, improving the environmental adaptability of edge recognition technology, reducing the number of parameters, and improving the computational efficiency, making edge recognition simpler, more convenient, and faster.
[0097] Secondly, a description is given of an edge detection device based on the material characteristics of self-sensitized neurons proposed by the embodiments of the present application with reference to the accompanying drawings.
[0098] Figure 7 is a block diagram of an edge detection device based on the material characteristics of self-sensitized neurons according to the embodiments of the present application.
[0099] As Figure 7 shown, the edge detection device 10 based on the material characteristics of self-sensitized neurons includes: a decomposition module 100, a calculation module 200, and a visualization module 300.
[0100] Among them, the decomposition module 100 is configured to read the original image data of the target and decompose the RGB channel image data in the original image data.
[0101] The calculation module 200 is configured to calculate the gradient information of the RGB channel image data and obtain the gradient information of the original image data based on the gradient information of the RGB channel image data.
[0102] A visualization module 300, configured to repeatedly input gradient information of original image data into a preset self-sensitizing neuron within a preset simulation time, adjust the state of the self-sensitizing neuron, calculate the pulse firing frequency of the self-sensitizing neuron, and convert the pulse firing frequency into grayscale pixel values, so as to perform visualization of preset target edge detection based on the grayscale pixel values.
[0103] Optionally, in an embodiment of the present application, the calculation module 200 includes: a smoothing unit, a first operation unit, and a second operation unit.
[0104] Wherein, the smoothing unit is configured to perform Gaussian smoothing operation on RGB channel image data to obtain an RGB channel Gaussian blurred image.
[0105] The first operation unit is configured to calculate the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image.
[0106] The second operation unit is configured to sum the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image to obtain the gradient information of the original image data of the preset target.
[0107] Optionally, in an embodiment of the present application, the edge detection device 10 based on the material characteristics of the self-sensitizing neuron of the embodiment of the present application further includes: a preprocessing module, an updating module, and a determination module.
[0108] Wherein, the preprocessing module is configured to preprocess the gradient information of the original image data before adjusting the state of the self-sensitizing neuron and calculating the pulse firing frequency of the self-sensitizing neuron to obtain the input data of the self-sensitizing neuron.
[0109] The updating module is configured to repeatedly input the input data into the self-sensitizing neuron within a preset simulation time, and update the membrane potential of the self-sensitizing neuron based on a preset self-sensitizing neuron potential calculation formula to obtain the latest neuron potential.
[0110] The determination module is configured to input the latest neuron potential into a preset neuron firing function to determine whether the self-sensitizing neuron fires.
[0111] Optionally, in an embodiment of the present application, the visualization module 300 includes: a judgment unit, an analysis unit, and a switching unit.
[0112] Wherein, the judgment unit is configured to, when the self-sensitizing neuron fires, obtain the historical firing pulse number of the self-sensitizing neuron, update the preset simulation time to obtain a new simulation time, and determine whether the new simulation time is a preset checkpoint.
[0113] An analysis unit, configured to determine whether a preset checkpoint and a historical excitation record meet a preset change condition if a new simulation time is the preset checkpoint.
[0114] A switching unit, configured to switch the state of the self-sensitizing neuron to a preset state if the preset change condition is met.
[0115] Optionally, in an embodiment of the present application, the visualization module 300 further includes: a recording unit and a third operation unit.
[0116] Wherein, the recording unit is configured to record the number of pulses of the self-sensitizing neuron.
[0117] The third operation unit is configured to calculate the pulse excitation frequency of the self-sensitizing neuron according to the number of pulses and the preset simulation time.
[0118] It should be noted that the foregoing explanation of the embodiment of the edge detection method based on the material characteristics of the self-sensitizing neuron also applies to the edge detection device based on the material characteristics of the self-sensitizing neuron in this embodiment, and will not be elaborated here.
[0119] An edge detection device based on the material characteristics of the self-sensitizing neuron according to an embodiment of the present application includes a decomposition module, configured to read the original image data of the target and decompose the RGB channel image data in the original image data; a calculation module, configured to calculate the gradient information of the RGB channel image data and obtain the gradient information of the original image data based on the gradient information of the RGB channel image data; a visualization module, configured to repeatedly input the gradient information of the original image data into a preset self-sensitizing neuron within a preset simulation time, adjust the state of the self-sensitizing neuron, calculate the pulse excitation frequency of the self-sensitizing neuron, and convert the pulse excitation frequency into a grayscale pixel value, so as to perform visualization of the preset target edge detection according to the grayscale pixel value, thereby breaking through the limitations of the application scenarios of traditional operators, improving the environmental adaptability of the edge recognition technology, reducing the number of parameters, improving the calculation efficiency, and making the edge recognition simpler, more convenient and faster.
[0120] Figure 8 The structural schematic diagram of the electronic device provided by the embodiment of the present application. The electronic device may include:
[0121] A memory 801, a processor 802, and a computer program stored on the memory 801 and executable on the processor 802.
[0122] When the processor 802 executes the program, it implements the edge detection method based on the material characteristics of the self-sensitizing neuron provided in the above embodiment.
[0123] Further, the electronic device further includes:
[0124] A communication interface 803 for communication between the memory 801 and the processor 802.
[0125] A memory 801 for storing a computer program that can run on the processor 802.
[0126] The memory 801 may include a high-speed RAM memory and may also include non-volatile memory, such as at least one disk memory.
[0127] If the memory 801, the processor 802, and the communication interface 803 are implemented independently, the communication interface 803, the memory 801, and the processor 802 can be interconnected via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 8 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.
[0128] Optionally, in a specific implementation, if the memory 801, the processor 802, and the communication interface 803 are integrated on a single chip, the memory 801, the processor 802, and the communication interface 803 can communicate with each other through an internal interface.
[0129] The processor 802 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.
[0130] The embodiments of the present application also provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the edge detection method based on the characteristics of the self-sensitizing neuron material as described above is implemented.
[0131] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.
[0132] In addition, the terms "first" and "second" are used only for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of this application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically defined.
[0133] Any process or method description shown in a flowchart or described in other ways herein can be understood to represent a module, segment, or portion of code including one or N executable instructions for implementing a customized logical function or process, and the scope of the preferred embodiments of this application includes additional implementations, where the functions may be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of this application belong.
[0134] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, which can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection part (electronic device) having one or N wirings, a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then stored in a computer memory.
[0135] It should be understood that various parts of the present application can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0136] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of implementing the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.
[0137] In addition, each functional unit in various embodiments of the present application may be integrated into a processing module, may exist separately as individual physical units, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
[0138] The above-mentioned storage medium may be a read-only memory, a magnetic disk, an optical disc, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.
Claims
1. An edge detection method based on the characteristics of self-sensitized neuron materials, characterized in that including the following steps: Reading the original image data of the target and decomposing the RGB channel image data in the original image data; Calculating the gradient information of the RGB channel image data, and based on the gradient information of the RGB channel image data, obtaining the gradient information of the original image data, and Within a preset simulation time, repeatedly inputting the gradient information of the original image data into a preset self-sensitizing neuron, adjusting the state of the self-sensitizing neuron, calculating the pulse firing frequency of the self-sensitizing neuron, and converting the pulse firing frequency into a grayscale pixel value to perform visualization of the preset target edge detection according to the grayscale pixel value.
2. The method according to claim 1, wherein The calculating the gradient information of the RGB channel image data and obtaining the gradient information of the original image data based on the gradient information of the RGB channel image data includes: Performing a Gaussian smoothing operation on the RGB channel image data to obtain an RGB channel Gaussian blurred image; Calculating the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image; Adding up the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image to obtain the gradient information of the original image data of the preset target.
3. The method according to claim 2, wherein Before adjusting the state of the self-sensitizing neuron and calculating the pulse firing frequency of the self-sensitizing neuron, it further includes: Preprocessing the gradient information of the original image data to obtain the input data of the self-sensitizing neuron; Within the preset simulation time, repeatedly inputting the input data into the self-sensitizing neuron and updating the membrane potential of the self-sensitizing neuron based on a preset self-sensitizing neuron potential calculation formula to obtain the latest neuron potential; Inputting the latest neuron potential into a preset neuron firing function to determine whether the self-sensitizing neuron fires.
4. The method according to claim 3, characterized in that, The adjusting the state of the self-sensitizing neuron includes: When the self-sensitizing neuron fires, obtaining the historical firing pulse number of the self-sensitizing neuron, updating the preset simulation time to obtain a new simulation time, and determining whether the new simulation time is a preset checkpoint; If the new simulation time is the preset checkpoint, determining whether the preset checkpoint and the historical firing record satisfy a preset change condition; If the preset change condition is satisfied, switching the state of the self-sensitizing neuron to a preset state.
5. The method according to claim 1, characterized in that, The calculating the pulse firing frequency of the self-sensitizing neuron includes: Recording the pulse number of the self-sensitizing neuron; Calculating the pulse firing frequency of the self-sensitizing neuron according to the pulse number and the preset simulation time.
6. An edge detection device based on the characteristics of self-sensitizing neuron materials, characterized in that, including: A decomposition module for reading the original image data of the target and decomposing the RGB channel image data in the original image data; A calculation module for calculating the gradient information of the RGB channel image data and obtaining the gradient information of the original image data based on the gradient information of the RGB channel image data, and A visualization module, configured to repeatedly input the gradient information of the original image data into a preset self-sensitizing neuron within a preset simulation time, adjust the state of the self-sensitizing neuron, calculate the pulse firing frequency of the self-sensitizing neuron, and convert the pulse firing frequency into grayscale pixel values, so as to perform visualization of the preset target edge detection according to the grayscale pixel values.
7. The device according to claim 6, characterized in that, The calculation module includes: A smoothing unit, configured to perform Gaussian smoothing operation on the RGB channel image data to obtain an RGB channel Gaussian blurred image; A first operation unit, configured to calculate the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image; A second operation unit, configured to sum up the gradient magnitude and direction of each pixel in the RGB channel Gaussian blurred image to obtain the gradient information of the original image data of the preset target.
8. The device according to claim 7, characterized in that, It further includes: A preprocessing module, configured to preprocess the gradient information of the original image data before adjusting the state of the self-sensitizing neuron and calculating the pulse firing frequency of the self-sensitizing neuron to obtain the input data of the self-sensitizing neuron; An updating module, configured to repeatedly input the input data into the self-sensitizing neuron within the preset simulation time and update the membrane potential of the self-sensitizing neuron based on a preset self-sensitizing neuron potential calculation formula to obtain the latest neuron potential; A determination module, configured to input the latest neuron potential into a preset neuron firing function to determine whether the self-sensitizing neuron fires.
9. An electronic device, characterized in that, It includes: A memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the edge detection method based on the material characteristics of the self-sensitizing neuron according to any one of claims 1-5.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to be used to implement the edge detection method based on the material characteristics of the self-sensitizing neuron according to any one of claims 1-5.