Non-bird false touch gating dynamic protection method based on temperature difference and displacement double threshold fusion

Abstract

The application relates to the field of gated dynamic protection. A non-bird touch gating dynamic protection method based on temperature difference and displacement double threshold fusion comprises the following steps: acquiring real-time monitoring data of a gating area, wherein the real-time monitoring data comprises a temperature difference characteristic parameter, a displacement characteristic parameter, an infrared shielding parameter and a visual image parameter; acquiring a temperature difference discrimination value according to the temperature difference characteristic parameter; acquiring a displacement discrimination value according to the displacement characteristic parameter; acquiring an infrared discrimination value according to the infrared shielding parameter; acquiring a visual discrimination value according to the visual image parameter; acquiring a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value; and judging whether the comprehensive discrimination index exceeds a preset double threshold decision interval. Through the cooperative analysis of multi-dimensional data such as temperature difference, displacement, infrared shielding and visual image, the application effectively distinguishes biological and non-biological contact and improves the detection accuracy in a complex scene.

Description

TECHNICAL FIELD

[0001] The application relates to the field of gate dynamic protection, in particular to a non-bird false touch gate dynamic protection method based on temperature difference and displacement double threshold fusion. BACKGROUND

[0002] The intelligent gate system is a typical application of the Internet of Things technology in the security field, and has been widely used in residential, commercial, industrial and other scenarios.

[0003] In the related art, the existing technology relies on infrared pairs or matrix detection of shielding, but cannot distinguish between biological and non-biological shielding (such as fallen leaves and dust instantaneous shielding, which are easy to trigger false alarms), and in strong light environments (such as noon in summer), the infrared sensor is easily disturbed, leading to missed detection or misjudgment, and there is a lack of multi-dimensional data fusion, and only a single index decision is made by the shielding area or the duration, which cannot exclude non-biological false touch scenarios such as temporary flying of birds or accumulation of sundries. SUMMARY

[0004] The application provides a non-bird false touch gate dynamic protection method based on temperature difference and displacement double threshold fusion to solve the problem of lack of multi-dimensional data fusion in the related art, only a single index decision is made by the shielding area or the duration, and non-biological false touch scenarios such as temporary flying of birds or accumulation of sundries cannot be excluded.

[0005] In a first aspect, the application provides a non-bird false touch gate dynamic protection method based on temperature difference and displacement double threshold fusion, which comprises:

[0006] Obtaining real-time monitoring data of a gate area, wherein the real-time monitoring data comprises temperature difference characteristic parameters, displacement characteristic parameters, infrared shielding parameters and visual image parameters;

[0007] Obtaining a temperature difference discrimination value according to the temperature difference characteristic parameters;

[0008] Obtaining a displacement discrimination value according to the displacement characteristic parameters;

[0009] Obtaining an infrared discrimination value according to the infrared shielding parameters;

[0010] Obtaining a visual discrimination value according to the visual image parameters;

[0011] Obtaining a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value;

[0012] Judging whether the comprehensive discrimination index exceeds a preset double threshold decision interval;

[0013] If the comprehensive discrimination index exceeds the preset double threshold decision interval, it is determined that it is a non-bird false touch and a hierarchical protection response is triggered;

[0014] If the comprehensive discrimination index does not exceed the preset double-threshold decision interval, it is determined that the state is normal and the original state of the gate is maintained.

[0015] Optionally, the step of obtaining a temperature difference discrimination value according to the temperature difference characteristic parameter comprises:

[0016] obtaining a real-time temperature difference sequence and an environmental reference temperature in the temperature difference characteristic parameter;

[0017] obtaining a plurality of adjacent time temperature difference difference values according to the real-time temperature difference sequence, and generating a temperature difference change rate sequence according to the plurality of adjacent time temperature difference difference values;

[0018] obtaining a temperature difference absolute value and a temperature difference duration according to the temperature difference change rate sequence and the environmental reference temperature;

[0019] obtaining a temperature difference-time change rate according to the temperature difference absolute value and the temperature difference duration;

[0020] mapping the temperature difference-time change rate to a preset interval to obtain a temperature difference discrimination value.

[0021] Optionally, the step of obtaining a displacement discrimination value according to the displacement characteristic parameter comprises:

[0022] obtaining a plurality of three-dimensional displacement coordinates and a timestamp sequence according to the displacement characteristic parameter;

[0023] obtaining a plurality of adjacent time displacement vector differences according to the plurality of three-dimensional displacement coordinates and the timestamp sequence, and obtaining a displacement velocity according to the adjacent time displacement vector differences;

[0024] obtaining a displacement distance according to the displacement velocity;

[0025] obtaining a displacement trajectory curve according to the displacement distance and the timestamp sequence;

[0026] obtaining a motion trajectory curvature according to the displacement trajectory curve;

[0027] obtaining a displacement discrimination value according to the displacement velocity, the displacement distance, and the motion trajectory curvature.

[0028] Optionally, the step of obtaining an infrared discrimination value according to the infrared shielding parameter comprises:

[0029] obtaining a plurality of shielding areas according to the infrared shielding parameter;

[0030] obtaining a total number of shielding pixels at a plurality of times according to the shielding area, and obtaining a shielding area area according to the total number of shielding pixels;

[0031] obtaining a shielding area pixel area proportion according to the shielding area area;

[0032] obtaining the occlusion duration according to the occlusion area pixel ratio;

[0033] obtaining the region direction change sequence according to the occlusion duration and the occlusion area pixel ratio;

[0034] obtaining the infrared discriminant value according to the region direction change sequence.

[0035] Optionally, the step of obtaining the visual discriminant value according to the visual image parameter comprises:

[0036] obtaining target bounding box coordinates and category picture information according to the visual image parameter;

[0037] obtaining a target aspect ratio according to the target bounding box coordinates;

[0038] obtaining a bounding box bottom coordinate according to the target bounding box coordinates;

[0039] obtaining a barycenter height according to the bounding box bottom coordinate and the target aspect ratio;

[0040] obtaining a category similarity from a preset library according to the category picture information;

[0041] obtaining the visual discriminant value according to the target aspect ratio, the barycenter height and the category similarity.

[0042] Optionally, the step of obtaining the comprehensive discriminant index according to the temperature difference discriminant value, the displacement discriminant value, the infrared discriminant value and the visual discriminant value comprises:

[0043] obtaining a first weight coefficient according to the temperature difference discriminant value;

[0044] obtaining a second weight coefficient according to the displacement discriminant value;

[0045] obtaining a third weight coefficient according to the infrared discriminant value;

[0046] obtaining a fourth weight coefficient according to the visual discriminant value;

[0047] obtaining an environment compensation coefficient;

[0048] obtaining the comprehensive discriminant index according to the temperature difference discriminant value, the displacement discriminant value, the infrared discriminant value, the visual discriminant value, the first weight coefficient, the second weight coefficient, the third weight coefficient, the fourth weight coefficient and the environment compensation coefficient.

[0049] In a second aspect, the present application provides a non-bird class false touch gating dynamic protection system based on temperature difference and displacement double threshold fusion, comprising:

[0050] The data acquisition module is configured to acquire real-time monitoring data of the gating area, wherein the real-time monitoring data comprises a temperature difference characteristic parameter, a displacement characteristic parameter, an infrared shielding parameter, and a visual image parameter.

[0051] The temperature difference analysis module is configured to acquire a temperature difference discrimination value according to the temperature difference characteristic parameter.

[0052] The displacement analysis module is configured to acquire a displacement discrimination value according to the displacement characteristic parameter.

[0053] The infrared analysis module is configured to acquire an infrared discrimination value according to the infrared shielding parameter.

[0054] The visual analysis module is configured to acquire a visual discrimination value according to the visual image parameter.

[0055] The fusion decision module is configured to acquire a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value, and the visual discrimination value.

[0056] The decision module is configured to determine whether the comprehensive discrimination index exceeds a preset double-threshold decision interval.

[0057] If the comprehensive discrimination index exceeds the preset double-threshold decision interval, it is determined that the non-bird species false touch occurs, and a hierarchical protection response is triggered.

[0058] If the comprehensive discrimination index does not exceed the preset double-threshold decision interval, it is determined that the normal state is maintained, and the original state of the gate is maintained.

[0059] Optionally, the temperature difference analysis module comprises:

[0060] The temperature difference sequence acquisition unit is configured to acquire a real-time temperature difference sequence and an environmental reference temperature in the temperature difference characteristic parameter.

[0061] The change rate generation unit is configured to acquire a plurality of adjacent time temperature difference difference values according to the real-time temperature difference sequence, and generate a temperature difference change rate sequence according to the plurality of adjacent time temperature difference difference values.

[0062] The temperature difference parameter calculation unit is configured to acquire a temperature difference absolute value and a temperature difference duration according to the temperature difference change rate sequence and the environmental reference temperature.

[0063] The temperature difference-duration change rate calculation unit is configured to acquire a temperature difference-duration change rate according to the temperature difference absolute value and the temperature difference duration.

[0064] The mapping unit is configured to map the temperature difference-duration change rate to a preset interval to obtain a temperature difference discrimination value.

[0065] Optionally, the displacement analysis module comprises:

[0066] The coordinate sequence acquisition unit is configured to acquire a plurality of three-dimensional displacement coordinates and a timestamp sequence according to the displacement feature parameter;

[0067] The speed calculation unit is configured to acquire a displacement vector difference between adjacent time points according to the plurality of three-dimensional displacement coordinates and the timestamp sequence, and obtain a displacement speed according to the displacement vector difference between adjacent time points;

[0068] The distance calculation unit is configured to acquire a displacement distance according to the displacement speed;

[0069] The trajectory fitting unit is configured to acquire a displacement trajectory curve according to the displacement distance and the timestamp sequence;

[0070] The curvature calculation unit is configured to acquire a motion trajectory curvature according to the displacement trajectory curve;

[0071] The displacement discrimination value generation unit is configured to acquire a displacement discrimination value according to the displacement speed, the displacement distance, and the motion trajectory curvature.

[0072] Optionally, the infrared analysis module comprises:

[0073] The occlusion region identification unit is configured to acquire a plurality of occlusion regions according to the infrared occlusion parameter;

[0074] The pixel statistics unit is configured to acquire a total number of occlusion pixels at a plurality of time points according to the occlusion regions, and acquire an occlusion region area according to the total number of occlusion pixels;

[0075] The area proportion calculation unit is configured to acquire an occlusion region pixel area proportion according to the occlusion region area;

[0076] The duration calculation unit is configured to acquire an occlusion duration according to the occlusion region pixel area proportion;

[0077] The direction change analysis unit is configured to acquire a region direction change sequence according to the occlusion duration and the occlusion region pixel area proportion;

[0078] The infrared discrimination value generation unit is configured to acquire an infrared discrimination value according to the region direction change sequence.

[0079] Compared with the related art, the non-bird dynamic protection method based on the temperature difference and displacement double-threshold fusion provided in the present application at least has the following technical effects:

[0080] Through the collaborative analysis of multi-dimensional data such as temperature difference, displacement, infrared shielding, visual image, etc., biological and non-biological contact can be effectively distinguished, and the detection accuracy in complex scenes can be improved; an environment self-adaptive mechanism is introduced, and the weight coefficient and compensation factor are dynamically adjusted according to real-time temperature, humidity, illumination and other parameters, so as to enhance the stability of the system in extreme environments; a bird-specific exclusion algorithm based on target morphological features and motion patterns is introduced to reduce natural factor interference and reduce false alarm rate; through a hierarchical response strategy, multi-level protection from warning to locking is realized, and safety and passing efficiency are balanced.

[0081] The details of one or more embodiments of the present application are presented in the following drawings and description to make other features, objects and advantages of the present application more apparent. BRIEF DESCRIPTION OF DRAWINGS

[0082] The drawings described herein are intended to provide further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their description serve to explain the present application, and do not constitute an improper limitation on the present application. In the drawings:

[0083] Figure 1 is a non-bird false touch gating dynamic protection method flow chart based on temperature difference and displacement double threshold fusion according to an exemplary embodiment;

[0084] Figure 2 is a system diagram of a non-bird false touch gating dynamic protection method based on temperature difference and displacement double threshold fusion according to an exemplary embodiment. DETAILED DESCRIPTION

[0085] In order to make the purpose, technical scheme and advantages of the present application more apparent, the present application is described and explained below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application. Based on the embodiments provided by the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0086] Obviously, the drawings in the following description are only some examples or embodiments of the present application, and for those of ordinary skill in the art, the present application can be applied to other similar scenarios without creative labor on the basis of these drawings. In addition, it can be understood that although the efforts made in this development process may be complex and lengthy, for those of ordinary skill in the art related to the content disclosed in the present application, some design, manufacture or production changes based on the technical content disclosed in the present application are only routine technical means, and should not be understood as insufficient disclosure of the present application.

[0087] Reference to an "embodiment" in this application means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. It is expressly understood that that the embodiments described herein are merely examples from a multitude of possible embodiments, which have been presented for illustrative purposes only. Other embodiments can be devised without departing from the spirit of the application.

[0088] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The materials, methods, and examples provided herein are illustrative only and not intended to be limiting. Except to the extent necessary or inherent in the processes themselves, capital terms such as "comprise", "comprises", "comprising", "include", "includes", "including", "contain", "contains", "containing", "have", "has", "having", "may", "might", "must", "need", "or", "shall", "shalls", "should", "shouldn't", "will", and "would" are not intended to be limiting. The term "a" or "an" means "one or more". The term "another" means at least a second or more. The terms "including", "comprising", "having" and variations thereof mean inclusion without limitation. For example, a process, method, article, or apparatus that "comprises" a list of steps or components herein can not necessarily be limited to those steps or components, but can include additional steps or components not expressly listed or inherent to such process, method, article, or apparatus. The words "plurality" and "a plurality" mean "two or more". Words using the singular or plural number also include the plural or singular number respectively. The word "coupled" means two or more separate elements or devices that are coupled together either directly or indirectly and can encompass a wired or wireless connection. The term "comprising", used in the detailed description and throughout the claims, is not intended to be limiting. The term "first", "second", "third", and so on, used in the detailed description and throughout the claims, is not intended to be limiting.

[0089] Embodiment 1

[0090] The embodiment of the application provides a non-bird touch-misoperation dynamic protection method based on temperature difference and displacement double threshold fusion. Figure 1 is a method flowchart shown according to an exemplary embodiment. As shown in Figure 1 , the method comprises:

[0091] Step S101, acquiring real-time monitoring data of a gating region, wherein the real-time monitoring data comprises a temperature difference feature parameter, a displacement feature parameter, an infrared shielding parameter and a visual image parameter;

[0092] Step S102, acquiring a temperature difference discrimination value according to the temperature difference feature parameter;

[0093] Step S103, acquiring a displacement discrimination value according to the displacement feature parameter;

[0094] Step S104, acquiring an infrared discrimination value according to the infrared shielding parameter;

[0095] Step S105, acquiring a visual discrimination value according to the visual image parameter;

[0096] Step S106, acquiring a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value;

[0097] Step S107, judging whether the comprehensive discrimination index exceeds a preset double-threshold decision interval;

[0098] If the comprehensive discrimination index exceeds the preset double-threshold decision interval, it is determined as a non-bird class false touch and a hierarchical protection response is triggered;

[0099] If the comprehensive discrimination index does not exceed the preset double-threshold decision interval, it is determined as a normal state and the original state of the gate is maintained.

[0100] In summary, the non-bird false touch dynamic protection method based on temperature difference and displacement double threshold fusion provided by the embodiment of the present application synchronously collects the temperature difference characteristic parameters (including real-time temperature difference sequence, environmental reference temperature), displacement characteristic parameters (including three-dimensional displacement coordinates, timestamp sequence), infrared shielding parameters (including shielding area, total number of shielding pixels), visual image parameters (including target bounding box coordinates, category picture information) and other real-time monitoring data of the door control area through the pyroelectric sensor, MEMS accelerometer, 16*16 infrared matrix, low-illumination camera and other devices deployed around the door, and then performs deep processing on each dimension of data: calculates the temperature difference change rate sequence from the temperature difference characteristic parameters by calculating the temperature difference difference value of adjacent moments, obtains the temperature difference absolute value and duration from the environmental reference temperature, and maps the temperature difference-time duration change rate to a preset interval to obtain a temperature difference discrimination value; obtains the displacement speed from the displacement characteristic parameters by calculating the displacement vector difference of adjacent moments, integrates to obtain the displacement distance, fits the displacement trajectory curve to calculate the curvature, and obtains the displacement discrimination value by weighting the three; obtains the area ratio from the infrared shielding parameters by counting the total number of shielding pixels, calculates the shielding duration and the entropy value of the area direction change sequence, and obtains the infrared discrimination value through formula calculation; obtains the target aspect ratio and barycentric height from the visual image parameters, obtains the category similarity by comparing with the preset library, calculates the visual discrimination value according to whether it meets the bird characteristic interval, and then combines the first to fourth weight coefficients determined according to the temperature difference, displacement, infrared and visual discrimination values and the environmental compensation coefficient to calculate the comprehensive discrimination index through the weighted fusion formula, for example: compares the index with the double threshold decision interval formed by the preset primary threshold T1=0.6 and the secondary threshold T2=0.8, if the comprehensive discrimination index exceeds T2 and the target category similarity excludes birds, it is determined that it is a non-bird false touch, and the graded protection responses such as electronic lock locking and cloud alarm are triggered, if it does not exceed T1 and the target category similarity matches birds, it is determined that it is in a normal state and the original state of the door control is maintained, and when it is between T1 and T2, a level one response such as sound and light warning is triggered. The scheme constructs a complete logical chain covering data acquisition, feature acquisition, risk assessment and graded decision making through the spatio-temporal alignment and feature level fusion of multi-modal sensor data, the adaptive adjustment of the dynamic threshold according to the environmental temperature, the door body age and other parameters, and the bird exclusion mechanism based on the target morphological characteristics and behavior patterns, effectively solves the problems of weak anti-interference ability of single sensor, poor environmental adaptability of fixed threshold and high false judgment rate of birds and non-bird targets, and makes the finally obtained non-bird false touch dynamic protection method take into account the influence of complex environment and door body characteristics, and optimizes the effect of non-bird false touch detection.

[0101] In one embodiment, step S102 specifically includes:

[0102] Step S1021, obtaining the real-time temperature difference sequence and the environmental reference temperature in the temperature difference characteristic parameters;

[0103] Step S1022, obtaining a plurality of adjacent time temperature difference difference values according to the real-time temperature difference sequence, and generating a temperature difference change rate sequence according to the plurality of adjacent time temperature difference difference values;

[0104] Step S1023, obtaining a temperature difference absolute value and a temperature difference duration according to the temperature difference change rate sequence and the environment reference temperature;

[0105] Step S1024, obtaining a temperature difference-duration change rate according to the temperature difference absolute value and the temperature difference duration;

[0106] Step S1025, mapping the temperature difference-duration change rate to a preset interval to obtain a temperature difference discrimination value.

[0107] In step S1021, the "real-time temperature difference sequence" is derived from the real-time continuous sampling of the pyroelectric sensor on the temperature in the gated area. The sensor continuously obtains the temperature values of the environment and objects around the door at a fixed frequency, forming a temperature data sequence arranged in chronological order, which is used to dynamically reflect the temperature change in the gated area. The "environment reference temperature" is obtained by statistically analyzing the historical data of the real-time temperature difference sequence, taking the average temperature value in the past period of time as the reference value for judging whether the current temperature is abnormal. Its role is to provide a stable baseline for temperature difference analysis. The system usually updates the environment reference temperature regularly (e.g., every hour) to adapt to environmental changes. The two work together to preliminarily identify the nature of temperature fluctuations through comparison of real-time temperature and reference temperature.

[0108] In step S1022, the "adjacent time temperature difference difference value" is the absolute difference value of the current time temperature value and the previous time temperature value obtained by processing the real-time temperature difference sequence by time. This difference value reflects the temperature change amplitude in unit time. By taking the absolute value, the influence of temperature rising and falling direction is eliminated, and the analysis of change speed is focused. The "temperature difference change rate sequence" is a sequence formed by arranging a plurality of adjacent time temperature difference difference values in chronological order, which is used to show the evolution trend of temperature change speed over time. The calculation process is to traverse the real-time temperature difference sequence, calculate the temperature absolute difference value of adjacent two time points, and store it in the sequence. The role of this sequence is to analyze the fast and slow patterns of temperature change, distinguish between rapid temperature fluctuations (such as body temperature radiation changes caused by rapid movement of living organisms) and slow temperature fluctuations (such as natural environmental warming or cooling), and provide a speed dimension basis for subsequent judgment.

[0109] The "absolute value of temperature difference" in step S1023 is the absolute value of the difference between the real-time temperature value and the ambient reference temperature value, which is used to measure the degree of deviation of the current temperature from the normal reference. The larger the absolute value, the more significant the temperature anomaly. The "duration of temperature difference" refers to the continuous duration of the absolute value of the temperature difference being greater than a preset threshold (e.g., 0.5°C set by the system), which is used to determine whether the temperature anomaly is a temporary disturbance or a persistent change. In the calculation, the system will monitor the absolute value of the temperature difference in real time. When it is detected that the absolute value exceeds the preset threshold, the timing function is started. When the absolute value of the temperature difference falls below the threshold, the timing is stopped, and the duration is obtained. The combination of the two can effectively filter out temporary temperature difference disturbances caused by factors such as gusts and short-term changes in light, and lock in persistent temperature difference anomalies caused by persistent contact or stay of living organisms.

[0110] In step S1024, the "temperature difference-duration change rate" is obtained by performing an arithmetic multiplication operation on the absolute value of the temperature difference and the duration of the temperature difference. This index combines the degree of temperature deviation from the reference and the duration of the anomaly, and aims to comprehensively evaluate the contribution of the temperature difference anomaly to the risk of false touch. The calculation logic is as follows: the larger the absolute value of the temperature difference, the more significant the temperature anomaly; the longer the duration, the more stable the anomaly state. The product of the two can more comprehensively reflect the potential false touch possibility, and the larger the value usually means the higher the possibility of non-avian objects or living organisms contacting the door body. It is an important intermediate parameter that connects single-dimensional temperature analysis and comprehensive risk assessment.

[0111] In step S1025, the "preset interval" is a numerical range set by the system to measure the false touch risk level corresponding to the temperature difference-duration change rate, usually in the interval [0, 1], which facilitates unified fusion calculation with the discrimination values of other dimensions. The "mapping" process is to linearly convert the actual value range of the temperature difference-duration change rate into the preset interval by normalization and other mathematical methods to obtain the temperature discrimination value. In the specific calculation, the system will determine the minimum and maximum values of the temperature difference-duration change rate based on historical data statistics (e.g., the minimum value in a normal environment is 0, and the maximum value in a typical false touch scenario is a certain specific value). Then, the corresponding proportion is calculated using the formula (temperature difference-duration change rate - minimum value) ÷ (maximum value - minimum value), and the temperature discrimination value located in the preset interval is obtained. This discrimination value converts the temperature-related physical quantity into a risk probability value with a unified dimension, which facilitates subsequent weighted fusion with the discrimination values of displacement, infrared, vision, and other dimensions to form a comprehensive discrimination index, providing a standardized temperature dimension basis for the final false touch judgment.

[0112] Continuing to refer to Figure 1 After step S102, step S103 is performed, as follows:

[0113] Step S1031: Acquire multiple three-dimensional displacement coordinates and time stamp sequences according to the displacement characteristic parameters;

[0114] Step S1032: obtaining displacement vector differences at adjacent moments according to the plurality of three-dimensional displacement coordinates and the timestamp sequence, and obtaining displacement velocity according to the displacement vector differences at adjacent moments;

[0115] Step S1033: obtaining a displacement distance according to the displacement speed;

[0116] Step S1034: obtaining a displacement trajectory curve graph according to the displacement distance and timestamp sequence;

[0117] Step S1035: obtaining the motion trajectory curvature according to the displacement trajectory curve graph;

[0118] Step S1036: Obtain a displacement discrimination value according to the displacement speed, displacement distance, and motion trajectory curvature.

[0119] In step S1031, the "multiple three-dimensional displacement coordinates" are collected collaboratively by a MEMS accelerometer and a laser displacement meter deployed on the door. The MEMS accelerometer infers displacement by sensing changes in acceleration, while the laser displacement meter directly acquires distance data through laser ranging. The two are combined to generate the object's position coordinates in the X, Y, and Z axes. Each coordinate is associated with a precise "timestamp" to mark the moment of acquisition, with millisecond-level accuracy, ensuring the temporal sequence of the displacement data. This data is used to construct the spatiotemporal trajectory of the object within the gated area, providing raw data support for subsequent motion feature analysis. The system continuously collects and stores coordinates and timestamps at a fixed frequency (e.g., 50 Hz), forming a continuous displacement-time dataset.

[0120] In step S1032, the “adjacent time displacement vector difference” is calculated by performing vector subtraction on the three-dimensional displacement coordinates corresponding to adjacent time stamps, that is, calculating the coordinate difference in the X, Y, and Z axis directions respectively ( , ), the displacement vector change from the previous moment to the current moment is obtained; the "displacement speed" is the modulus of the vector difference divided by the time interval, and the calculation formula is , where Δt is the time interval between adjacent timestamps. This process converts spatial displacement into a velocity indicator, which can be used to measure the speed and direction of an object's movement. For example, a high-frequency, small-amplitude vector difference might correspond to the rapid micro-movement of a bird's wingbeat, while a low-frequency, large-amplitude vector difference might correspond to the movement of a human or pet.

[0121] In step S1033, the “displacement distance” is obtained by time-integrating the displacement velocity, i.e., the cumulative value of the product of the velocity and the time interval from the initial moment to the current moment ( ). This index is used to reflect the total moving path length of the object within the gating region, and its role is to assist in judging the continuity and scale of the movement: short distance displacement may be caused by false touch or small animal activity, and long distance displacement is more likely to be normal passage or large object movement. The system updates the displacement distance value in real time by dynamically accumulating the speed increment in each time interval, forming a cumulative quantity that grows over time.

[0122] In step S1034, the "displacement trajectory graph" uses an interpolation algorithm (such as cubic spline interpolation) to continuously fit the discrete displacement distance data with time stamp as the horizontal axis and displacement distance as the vertical axis, generating a smooth curve. The role of this graph is to visually show the trend of the distance of the object movement changing over time, making it easy to analyze the morphological characteristics of the trajectory, such as bird flight may present an irregular fluctuation curve, while human walking or object hitting the door may present a smooth straight line or low-frequency fluctuation curve. By visualizing the trajectory shape, the system can further obtain the geometric characteristics of the trajectory.

[0123] In step S1035, the "motion trajectory curvature" is calculated by differential geometry method, specifically, the second derivative of the displacement trajectory curve is calculated to obtain the acceleration, and the bending degree of the curve is calculated combined with the first derivative (speed), the formula is , where K represents the motion trajectory curvature, S represents the displacement distance, and d represents the displacement speed. The larger the curvature value, the higher the bending degree of the trajectory, and vice versa, which is closer to a straight line. Its significance lies in distinguishing the behavior patterns of different moving subjects: birds have high trajectory curvature due to strong flight maneuverability; the motion trajectory curvature of humans, pets or objects is relatively low, especially when moving in a straight line, the curvature tends to zero. This index provides a morphological basis for judging whether the movement is caused by non-avian organisms.

[0124] In step S1036, the "displacement discriminant value" is obtained by weighted fusion of the displacement speed, displacement distance, and motion trajectory curvature, the calculation formula is e= , where is the weight coefficient, which is set according to the contribution of each index to the false touch risk (such as the sensitivity weight of speed to fast movement is higher, and the weight of curvature to motion pattern is second). The role of this process is to convert multi-dimensional displacement features into a unified risk quantization index: the faster the speed, the longer the distance, the lower the curvature (the closer to a straight line), the higher the displacement discriminant value, indicating a higher risk of non-avian false touch; on the contrary, if the speed is slow, the distance is short, and the curvature is high, it may belong to birds or natural interference, and the risk is low. The system generates a dynamic displacement discriminant value by calculating the weighted sum of the three indexes in real time, providing a key parameter in the displacement dimension for comprehensive decision-making.

[0125] Continuing to refer to Figure 1 , step S104 is performed after step S103, specifically as follows:

[0126] Step S1041, obtaining a plurality of shielding areas according to the infrared shielding parameter;

[0127] Step S1042, obtaining a total number of shielding pixels at a plurality of moments according to the shielding areas, and obtaining a shielding area according to the total number of shielding pixels;

[0128] Step S1043, obtaining a shielding area pixel area proportion according to the shielding area;

[0129] Step S1044, obtaining a shielding duration according to the shielding area pixel area proportion;

[0130] Step S1045, obtaining a region direction change sequence according to the shielding duration and the shielding area pixel area proportion;

[0131] Step S1046, obtaining an infrared discrimination value according to the region direction change sequence.

[0132] In step S1041, the "infrared shielding parameter" is derived from a 16x16 infrared matrix sensor installed on the door body. The sensor uses a grid-shaped distribution of independent detection units to real-time perceive the object shielding condition in the door control area. Each unit outputs a binary signal (1 represents shielding, and 0 represents no shielding). The "plurality of shielding areas" is a continuous shielding unit cluster identified by image processing algorithm on the binary matrix output by the sensor, which is used to determine the specific spatial position and range of shielding occurrence. Its significance lies in converting the shielding state of the infrared signal into analyzable spatial distribution data, providing a basis for subsequent quantitative analysis. The calculation process is that the sensor generates a binary matrix by scanning at a fixed period. The system obtains independent shielding areas through algorithms such as contour detection and connected component labeling.

[0133] In step S1042, the "total number of shielding pixels" is the result of counting the number of sensor units that are shielded in each shielding area. It is obtained by iterating all units in the area and accumulating the number of pixels with a state of 1. The "shielding area" is represented by this statistical value, which is used to quantify the coverage scale of the shielding event in space. Its role is to distinguish between small shielding (such as insects flying by) and large shielding (such as biological body contact). The calculation process is to perform a summation operation on the binary matrix of each shielding area. The numerical value directly reflects the physical range of the shielding. For example, if a shielding area contains a number of continuous units, the total number of pixels in that area is the quantitative value of the area.

[0134] The "obstruction area pixel area proportion" in step S1043 is a proportional value obtained by dividing the total number of obstruction pixels by the total number of pixels of the infrared matrix (16x16=256), and the value range is 0 to 1. The "obstruction area pixel area proportion" is used to convert different scales of obstruction areas into a standardized index of uniform dimension, eliminate the influence of sensor grid resolution, and facilitate the comparison of obstruction intensity across time and across scenes. The calculation process is a simple division operation. For example, when the total number of obstruction pixels is a certain value, the proportion is obtained by dividing 256. The index can directly reflect the relative size of the obstruction area in the entire monitoring range.

[0135] In step S1044, the "obstruction duration" is obtained by monitoring the pixel area proportion curve of the obstruction area in real time. When the proportion continuously exceeds the preset threshold value (such as the minimum effective obstruction proportion set by the system), the timing function is started and the time difference from the first time the threshold value is exceeded to the first time the threshold value is lowered is recorded. The "obstruction duration" is used to distinguish between transient obstruction (such as the moment when a leaf falls) and persistent obstruction (such as an object leaning against the door). The role of the "obstruction duration" is to filter obstruction events with actual risks in the time dimension. The calculation process is to dynamically track the change of the area proportion. When the value continuously meets the condition, the timing starts and stops, and the time length record of the effective obstruction is formed.

[0136] In step S1045, the "area direction change sequence" is determined by calculating the centroid coordinate change of the obstruction area at consecutive time points. The centroid coordinates are calculated by the weighted average method (such as multiplying the coordinates of each obstruction unit by its state value, summing up, and then dividing by the total number of obstruction pixels). The vector difference of the centroid coordinates of adjacent time points is used to determine the direction (such as Δx>0 for right movement and Δy>0 for upward movement). A direction sequence arranged in time order is formed (such as "left movement-stationary-right movement"). The significance of the "area direction change sequence" is to capture the motion pattern of the obstruction. Birds flying are often accompanied by high-frequency direction changes, while non-bird false touches (such as pets scratching the door) are mostly one-way or low-frequency reciprocating motion. The calculation process involves dynamic updating of centroid coordinates and direction category division.

[0137] In step S1046, the "infrared discrimination value" is obtained by fusing the area proportion of the obstruction area, the obstruction duration, and the Shannon entropy value of the direction change sequence. First, the probability distribution of the direction sequence is calculated Then, the Shannon entropy formula is used to measure the uncertainty of direction change (the greater the entropy value, the more random the direction is. In the formula, H represents the modified entropy value, The final discrimination value is calculated by the formula "area ratio x duration x (1-entropy value / ln4)", wherein (1-entropy value / ln4) is used to suppress the influence of random motion (such as birds). Its role is to integrate the spatial coverage size, time duration and motion mode regularity into a single risk index, and the higher the value, the greater the possibility of non-birds being touched. The calculation process involves nonlinear combination of multidimensional data, realizing hierarchical conversion from original signal to risk discrimination.

[0138] With reference to the foregoing Figure 1 Step S105 is performed after step S104, and specifically as follows:

[0139] Step S1051, obtaining target bounding box coordinates and category picture information according to the visual image parameters;

[0140] Step S1052, obtaining a target aspect ratio according to the target bounding box coordinates;

[0141] Step S1053, obtaining a bounding box bottom coordinate according to the target bounding box coordinates;

[0142] Step S1054, obtaining a gravity center height according to the bounding box bottom coordinate and the target aspect ratio;

[0143] Step S1055, obtaining a category similarity from a preset library according to the category picture information;

[0144] Step S1056, obtaining a visual discrimination value according to the target aspect ratio, the gravity center height and the category similarity.

[0145] In step S1051, the "visual image parameters" are collected in real time by a low-illumination camera configured on the door body. The device can obtain the RGB image and depth data of the door control area under 0.01 lux illumination. The "target bounding box coordinates" are generated by the YOLOv5s lightweight target detection model for image inference. The model outputs the position coordinates of the target in the image (such as the upper left corner (x1, y1) and the lower right corner (x2, y2)). The "category picture information" is the target category label (such as "bird", "pet" and "person") and the corresponding feature vector (including color, texture, contour and other multidimensional visual features) predicted by the model. Its role is to convert visual information into quantifiable spatial position and category probability data. The calculation process is that the camera continuously collects image frames, and the model analyzes and identifies each frame to output the bounding box coordinates and the category confidence (such as a "bird" confidence of 0.8 indicating an 80% probability of being a bird).

[0146] In step S1052, the "target aspect ratio" is calculated by the target bounding box coordinates, and the formula is the ratio of the width to the height of the bounding box ), a metric used to describe the two-dimensional morphological characteristics of a target. Birds, due to their ellipsoidal shape, typically have an aspect ratio between 0.8 and 1.2. Non-avian targets (such as the rectangular silhouette of a standing human or the vertically elongated bodies of pets) often have aspect ratios greater than 1.5 or less than 0.8. This ratio is calculated by directly dividing the difference in bounding box coordinates to produce a dimensionless ratio. A value closer to 1 indicates a more avian-like morphology, while a lower value indicates a more non-avian morphological characteristic.

[0147] In step S1053, the "bounding box bottom coordinate" is the lowest row coordinate (y2) of the target bounding box in the image. This value must be physically calibrated based on the camera's installation height and viewing angle (e.g., by pre-measuring and mapping image pixels to actual distances) to convert pixel coordinates into actual physical heights. For example, if the camera is installed at a height of 2.5 meters and has a vertical viewing angle of 60 degrees, the ground height corresponding to the bottom pixel of the image can be calculated. A larger y2 bounding box bottom coordinate indicates a lower target height (e.g., a y2 closer to the bottom of the image indicates a target closer to the ground). This value is used to determine the target's vertical position. Birds typically move below 0.6 meters, while non-bird targets (such as standing people) often have a center of gravity above 0.8 meters.

[0148] In step S1054, the "center of gravity height" is estimated based on the bottom coordinates of the bounding box and the aspect ratio of the target. Assuming that the center of gravity of the target is located in the middle and lower part of the bounding box in the vertical direction (such as 1 / 3 of the height above the bottom), the calculation formula is center of gravity height = is the bounding box height. This estimation method is based on the distribution of the center of gravity of common targets (for example, birds have a lower center of gravity, while humans have a center of gravity approximately in the middle of the torso). The image coordinates of the center of gravity are calculated by scaling the bounding box height. This is then converted to an actual height using physical calibration. This helps determine whether the target is a low-altitude bird. For example, if the calculated center of gravity height is 0.5 meters, it is likely a bird; if it is 1.2 meters, it is more likely a person or a large pet.

[0149] In step S1055, "category similarity" is determined by comparing the target's visual feature vector with a bird feature template in a preset library. This library stores typical features (such as feather texture distribution and wing shape parameters) trained on a large number of bird images. A cosine similarity algorithm is used to calculate the degree of match between the target's features and the bird template, with the output value ranging from 0 to 1 (larger values ​​indicate greater similarity). This algorithm uses prior knowledge from machine learning to make a probabilistic judgment about the target's category. For example, a category similarity of 0.9 indicates a high degree of confidence that the target is a bird, thus directly eliminating false detections. A similarity of 0.3 indicates significant differences between the target and bird features, requiring further analysis based on other dimensional data.

[0150] In step S1056, the "visual discrimination value" is generated by weighted fusion of the target's aspect ratio, center of gravity height, and category similarity. H is the bird center of gravity height threshold (e.g., 0.6 meters), and C is the category similarity. The specific calculation logic is: ① The smaller the aspect ratio deviates from 1, the closer (1-|AR-1|) is to 1, indicating a more bird-like appearance. ② If the center of gravity height ≥ H, the object is considered non-bird, and the corresponding item is 1; otherwise, it is 0. ③ The lower the category similarity, the closer (1-C) is to 1, indicating a higher probability of non-bird contact. The weighted sum of these three items yields a discrimination value in the 0-1 range. Higher values ​​indicate a higher risk of mistouching a non-bird object.

[0151] Continue to refer to Figure 1 , after step S105, step S106 is performed, specifically as follows:

[0152] Step S1061: obtaining a first weight coefficient according to the temperature difference discrimination value;

[0153] Step S1062: obtaining a second weight coefficient according to the displacement discrimination value;

[0154] Step S1063: obtaining a third weight coefficient according to the infrared discrimination value;

[0155] Step S1064: obtaining a fourth weight coefficient according to the visual discrimination value;

[0156] Step S1065: Obtaining an environmental compensation coefficient;

[0157] Step S1066: Obtain a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value, the visual discrimination value, the first weight coefficient, the second weight coefficient, the third weight coefficient, the fourth weight coefficient and the environmental compensation coefficient.

[0158] The acquisition of the "first weight coefficient" in step S1061 of the present invention is based on a dynamic evaluation mechanism for temperature difference discrimination values. Its vocabulary originates from the reliability of temperature difference data in different environments, and its significance lies in reflecting the contribution of temperature difference characteristics to the discrimination of non-bird false touches. Specifically, the system pre-establishes a mapping relationship between temperature difference discrimination values ​​and weight coefficients. This relationship is derived from the statistical response sensitivity of temperature difference indicators to false touch events in historical data. For example, in low-temperature environments, temperature differences are more indicative of biological contact, and the corresponding first weight coefficient is automatically increased when the temperature difference discrimination value is higher. Its function is to dynamically adjust the weight of temperature difference data in the comprehensive decision-making process based on its credibility. During calculation, the corresponding weight is obtained by searching a preset nonlinear mapping table (such as a piecewise function or neural network model output), achieving adaptive weighting of the temperature difference dimension.

[0159] The second weight coefficient is associated with the displacement discriminant value in step S1062, the vocabulary of which is derived from the distinguishing ability of displacement features to motion patterns, and the meaning is to quantify the comprehensive influence of displacement speed, distance, curvature and other features on the risk of accidental touch. In the calculation process, the system adjusts the weight using an exponential or polynomial function according to the size of the displacement discriminant value, for example, when the displacement discriminant value is low (such as low-speed long-distance movement), it is considered to be normal traffic behavior, and the second weight coefficient is automatically reduced; when the discriminant value is high (such as high-speed short-distance vibration), it is determined to be a high-risk accidental touch, and the weight coefficient is correspondingly increased. The role is to highlight the indication of high-risk motion patterns in displacement data through a nonlinear weight distribution strategy, and to enhance the sensitivity to pet impact, object collision and other scenes.

[0160] The determination of the third weight coefficient in step S1063 depends on the infrared discriminant value, the vocabulary of which is related to the continuity and directionality of the infrared occlusion, and the meaning is to distinguish between effective occlusion and natural interference (such as falling leaves, gusts). The specific method is: a threshold value of the infrared discriminant value is preset, when the infrared discriminant value is greater than the threshold value (such as continuous and stable occlusion), the third weight coefficient takes a higher value, indicating that the occlusion event has a higher risk of accidental touch; when the discriminant value is less than the threshold value (such as instantaneous random occlusion), the weight coefficient takes a lower value. The role is to filter out invalid interference in infrared data through a step-by-step weight setting, and the weight assignment after threshold comparison is realized through a conditional judgment statement in the calculation, ensuring that the infrared dimension only plays an important role in the effective occlusion scene.

[0161] The fourth weight coefficient in step S1064 is determined by the visual discriminant value, the vocabulary of which is derived from the direct recognition ability of visual images to target categories, and the meaning is to prioritize the use of visual features to exclude bird targets. The calculation logic is: when the visual discriminant value is low (such as target category similarity matching birds), the fourth weight coefficient is automatically reduced, reducing the proportion of visual data in the comprehensive decision; when the discriminant value is high (such as the target being a pet or a person), the weight coefficient is increased, strengthening the dominant role of visual features. The role is to quickly filter out bird-related false positives through a category priority strategy, and in the calculation, a method of logical judgment combined with linear interpolation is used to dynamically assign weights according to the interval (such as [0, 0.5], (0.5, 1]) in which the visual discriminant value is located, realizing efficient differentiation between bird and non-bird targets.

[0162] The obtaining of the "environment compensation coefficient" in step S1065 is based on real-time environmental parameters, including temperature, light intensity, wind speed, etc., the vocabulary source of which is related to the influence of environmental factors on sensor performance, and the meaning is to eliminate the interference of environmental noise on each dimension data. The specific process is: real-time collection of parameters by environmental sensors (such as thermohygrograph, light sensor, anemometer) deployed in the door control area, calculation of compensation factors by using a pre-established compensation model, for example, the temperature difference sensor drift is smaller in high temperature environment, the compensation coefficient increases the weight of the temperature difference; the false positive rate of the infrared matrix increases in strong light environment, the compensation coefficient attenuates the weight of the infrared, dynamically calibrates the weight of each dimension, and ensures the stability of the system in complex environments such as -25℃~+60℃ wide temperature and 0.01 lux low light.

[0163] The calculation of the "comprehensive discrimination index" in step S1066 is the core link of multi-dimensional data fusion, and the process is: first, the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value are multiplied by the corresponding first to fourth weight coefficients respectively to obtain the weighted values of each dimension; then the weighted values are summed to form a preliminary comprehensive index; finally, the environmental compensation coefficient is multiplied to obtain the final comprehensive discrimination index. The calculation formula can be expressed as: , wherein, represents the temperature difference discrimination value, represents the displacement discrimination value, represents the infrared discrimination value, represents the visual discrimination value, represents the first weight coefficient, represents the second weight coefficient, represents the third weight coefficient, represents the fourth weight coefficient, and represents the environmental compensation coefficient. The meaning of this index is to convert the four types of heterogeneous data of temperature difference, displacement, infrared and vision into a unified dimension risk value, and the role is to provide a unique basis for subsequent double threshold decision. In the calculation process, the weight distribution of each dimension data reflects the principle of "morphological characteristics first, dynamic calibration of environment", for example, when the morphological characteristics of birds are obvious (the visual discrimination value is low), even if the discrimination values of other dimensions are high, the comprehensive index will be suppressed due to the decrease of the visual weight, avoiding misjudgment; when the non-bird target continuously contacts the door (the discrimination values of each dimension are high and the environmental compensation coefficient is increased), the comprehensive index quickly exceeds the threshold, triggering the graded protection response.

[0164] Embodiment 2

[0165] The embodiment 2 of the present application provides a non-bird door control dynamic protection system based on temperature difference and displacement double threshold fusion. Figure 2 is a system block diagram according to an exemplary embodiment. As Figure 2As shown, the non-bird false touch gating dynamic protection system based on temperature difference and displacement double threshold fusion comprises:

[0166] A data acquisition module 1 is configured to acquire real-time monitoring data of a gating area, wherein the real-time monitoring data comprises temperature difference characteristic parameters, displacement characteristic parameters, infrared shielding parameters and visual image parameters;

[0167] A temperature difference analysis module 2 is configured to acquire a temperature difference discrimination value according to the temperature difference characteristic parameters;

[0168] A displacement analysis module 3 is configured to acquire a displacement discrimination value according to the displacement characteristic parameters;

[0169] An infrared analysis module 4 is configured to acquire an infrared discrimination value according to the infrared shielding parameters;

[0170] A visual analysis module 5 is configured to acquire a visual discrimination value according to the visual image parameters;

[0171] A fusion decision module 6 is configured to acquire a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value;

[0172] A decision module 7 is configured to determine whether the comprehensive discrimination index exceeds a preset double threshold decision interval;

[0173] If the comprehensive discrimination index exceeds the preset double threshold decision interval, it is determined that it is a non-bird false touch and a hierarchical protection response is triggered;

[0174] If the comprehensive discrimination index does not exceed the preset double threshold decision interval, it is determined that it is a normal state and the original gating state is maintained.

[0175] In one embodiment, the temperature difference analysis module 2 comprises:

[0176] A temperature difference sequence acquisition unit is configured to acquire a real-time temperature difference sequence and an environmental reference temperature in the temperature difference characteristic parameters;

[0177] A change rate generation unit is configured to acquire a plurality of adjacent time temperature difference difference values according to the real-time temperature difference sequence, and generate a temperature difference change rate sequence according to the plurality of adjacent time temperature difference difference values;

[0178] A temperature difference parameter calculation unit is configured to acquire a temperature difference absolute value and a temperature difference duration according to the temperature difference change rate sequence and the environmental reference temperature;

[0179] A temperature difference-duration change rate calculation unit is configured to acquire a temperature difference-duration change rate according to the temperature difference absolute value and the temperature difference duration;

[0180] A mapping unit is configured to map the temperature difference-duration change rate to a preset interval to obtain a temperature difference discrimination value.

[0181] In one embodiment, the displacement analysis module 3 comprises:

[0182] a coordinate sequence acquisition unit configured to acquire a plurality of three-dimensional displacement coordinates and a timestamp sequence according to the displacement feature parameter;

[0183] a speed calculation unit configured to acquire a displacement vector difference between adjacent time instants according to the plurality of three-dimensional displacement coordinates and the timestamp sequence, and obtain a displacement speed according to the displacement vector difference between adjacent time instants;

[0184] a distance calculation unit configured to acquire a displacement distance according to the displacement speed;

[0185] a trajectory fitting unit configured to acquire a displacement trajectory curve according to the displacement distance and the timestamp sequence;

[0186] a curvature calculation unit configured to acquire a motion trajectory curvature according to the displacement trajectory curve;

[0187] a displacement discriminant value generation unit configured to acquire a displacement discriminant value according to the displacement speed, the displacement distance, and the motion trajectory curvature.

[0188] In one embodiment, the infrared analysis module 4 comprises:

[0189] an occlusion region identification unit configured to acquire a plurality of occlusion regions according to the infrared occlusion parameter;

[0190] a pixel statistics unit configured to acquire a total number of occlusion pixels at a plurality of time instants according to the occlusion regions, and acquire an occlusion region area according to the total number of occlusion pixels;

[0191] an area proportion calculation unit configured to acquire an occlusion region pixel area proportion according to the occlusion region area;

[0192] a duration calculation unit configured to acquire an occlusion duration according to the occlusion region pixel area proportion;

[0193] a direction change analysis unit configured to acquire a region direction change sequence according to the occlusion duration and the occlusion region pixel area proportion;

[0194] an infrared discriminant value generation unit configured to acquire an infrared discriminant value according to the region direction change sequence.

[0195] The above-described embodiments are merely illustrative of several embodiments of the present application, which are described in more detail and in a specific manner, but should not be construed as limiting the scope of the patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A non-bird touch-misoperation dynamic protection method based on temperature difference and displacement double threshold fusion, characterized in that, The method comprises: acquiring real-time monitoring data of a gating area, wherein the real-time monitoring data comprises a temperature difference characteristic parameter, a displacement characteristic parameter, an infrared shielding parameter and a visual image parameter; acquiring a real-time temperature difference sequence in the temperature difference characteristic parameter and an environmental reference temperature; acquiring a plurality of adjacent time temperature difference differences according to the real-time temperature difference sequence, and generating a temperature difference change rate sequence according to the plurality of adjacent time temperature difference differences; acquiring a temperature difference absolute value and a temperature difference duration according to the temperature difference change rate sequence and the environmental reference temperature; acquiring a temperature difference-time change rate according to the temperature difference absolute value and the temperature difference duration; mapping the temperature difference-time change rate to a preset interval to obtain a temperature difference discrimination value; acquiring a plurality of three-dimensional displacement coordinates and a timestamp sequence according to the displacement characteristic parameter; acquiring an adjacent time displacement vector difference according to the plurality of three-dimensional displacement coordinates and the timestamp sequence, and obtaining a displacement speed according to the adjacent time displacement vector difference; acquiring a displacement distance according to the displacement speed; acquiring a displacement trajectory curve according to the displacement distance and the timestamp sequence; acquiring a motion trajectory curvature according to the displacement trajectory curve; acquiring a displacement discrimination value according to the displacement speed, the displacement distance and the motion trajectory curvature; acquiring a plurality of shielding areas according to the infrared shielding parameter; acquiring a total number of shielding pixels at a plurality of times according to the shielding area, and acquiring a shielding area area according to the total number of shielding pixels; acquiring a shielding area pixel area proportion according to the shielding area area; acquiring a shielding duration according to the shielding area pixel area proportion; acquiring a region direction change sequence according to the shielding duration and the shielding area pixel area proportion; acquiring an infrared discrimination value according to the region direction change sequence; acquiring target bounding box coordinates and category picture information according to the visual image parameter; acquiring a target aspect ratio according to the target bounding box coordinates; acquiring a bounding box bottom coordinate according to the target bounding box coordinates; acquiring a barycenter height according to the bounding box bottom coordinate and the target aspect ratio; acquiring a category similarity from a preset library according to the category picture information; acquiring a visual discrimination value according to the target aspect ratio, the barycenter height and the category similarity; acquiring a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value; determining whether the comprehensive discrimination index exceeds a preset double-threshold decision interval; if the comprehensive discrimination index exceeds the preset double-threshold decision interval, determining that it is a non-bird class false touch and triggering a hierarchical protection response; if the comprehensive discrimination index does not exceed the preset double-threshold decision interval, determining that it is a normal state and maintaining the original state of the gating.

2. The dynamic protection method against false touch based on the fusion of the temperature difference and displacement double thresholds according to claim 1, characterized in that, The step of acquiring a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value comprises: acquiring a first weight coefficient according to the temperature difference discrimination value; acquiring a second weight coefficient according to the displacement discrimination value; acquiring a third weight coefficient according to the infrared discrimination value; acquiring a fourth weight coefficient according to the visual discrimination value; acquiring an environmental compensation coefficient; The comprehensive discrimination index is obtained according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value, the visual discrimination value, the first weight coefficient, the second weight coefficient, the third weight coefficient, the fourth weight coefficient and the environmental compensation coefficient.

3. A non-bird touch-misoperation dynamic protection system based on temperature difference and displacement double threshold fusion, characterized in that, Comprise: A data acquisition module is configured to acquire real-time monitoring data of a gated area, wherein the real-time monitoring data comprises a temperature difference characteristic parameter, a displacement characteristic parameter, an infrared shielding parameter and a visual image parameter; A temperature difference analysis module is configured to acquire a real-time temperature difference sequence in the temperature difference characteristic parameter and an environmental reference temperature; A plurality of adjacent time temperature difference differences are obtained according to the real-time temperature difference sequence, and a temperature difference change rate sequence is generated according to the plurality of adjacent time temperature difference differences; A temperature difference absolute value and a temperature difference duration are obtained according to the temperature difference change rate sequence and the environmental reference temperature; A temperature difference-duration change rate is obtained according to the temperature difference absolute value and the temperature difference duration; The temperature difference-duration change rate is mapped to a preset interval to obtain a temperature difference discrimination value; A displacement analysis module is configured to acquire a plurality of three-dimensional displacement coordinates and a timestamp sequence according to the displacement characteristic parameter; Adjacent time displacement vector differences are obtained according to the plurality of three-dimensional displacement coordinates and the timestamp sequence, and a displacement velocity is obtained according to the adjacent time displacement vector differences; A displacement distance is obtained according to the displacement velocity; A displacement trajectory curve graph is obtained according to the displacement distance and the timestamp sequence; A motion trajectory curvature is obtained according to the displacement trajectory curve graph; A displacement discrimination value is obtained according to the displacement velocity, the displacement distance and the motion trajectory curvature; An infrared analysis module is configured to acquire a plurality of shielding areas according to the infrared shielding parameter; A total number of shielding pixels at a plurality of times is obtained according to the shielding areas, and a shielding area area is obtained according to the total number of shielding pixels; A shielding area pixel area proportion is obtained according to the shielding area area; A shielding duration is obtained according to the shielding area pixel area proportion; A region direction change sequence is obtained according to the shielding duration and the shielding area pixel area proportion; An infrared discrimination value is obtained according to the region direction change sequence; A visual analysis module is configured to acquire target bounding box coordinates and category picture information according to the visual image parameter; A target aspect ratio is obtained according to the target bounding box coordinates; A bounding box bottom coordinate is obtained according to the target bounding box coordinates; A gravity center height is obtained according to the bounding box bottom coordinate and the target aspect ratio; A category similarity is obtained from a preset library according to the category picture information; A visual discrimination value is obtained according to the target aspect ratio, the gravity center height and the category similarity; A fusion decision module is configured to obtain a comprehensive discrimination index according to the temperature difference discrimination value, the displacement discrimination value, the infrared discrimination value and the visual discrimination value; A decision module is configured to determine whether the comprehensive discrimination index exceeds a preset double-threshold decision interval; If the comprehensive discrimination index exceeds the preset double-threshold decision interval, it is determined as a non-bird species false touch and a hierarchical protection response is triggered; If the comprehensive discrimination index does not exceed the preset double-threshold decision interval, it is determined as a normal state and the original state of the gate is maintained.

Images