A laser-vision-linked nighttime intrusion detection system

Abstract

The present invention discloses a laser-vision linkage nighttime intrusion detection system, which relates to the technical field of intrusion detection. The system includes a laser sensor, a vision sensor, a data processing module and a control module. The present invention utilizes the coordination between the various units in the control module to achieve the effect of automatically adjusting the sensor parameters according to the environment and the state of the target object, enhances the power of the laser and the sensitivity of the vision sensor in a weak light environment, increases the sampling frequency of the sensor when the target object moves quickly, adjusts the coordination mode of the laser and the vision sensor according to the distance of the target object, improves the accuracy and adaptability of detection, and integrates and analyzes the laser and vision data through the data processing module, and jointly establishes a dynamic linkage model with the control module, and utilizes the high-precision distance information of the laser data and the rich image features of the vision data to achieve high-precision detection and recognition of the target object.

Description

Technical Field

[0001] The present invention relates to the technical field of intrusion detection, and in particular to a laser-vision linked nighttime intrusion detection system. Background Art

[0002] In today's society, security protection is of vital importance. As one of the key means of security protection, the intrusion detection system mainly monitors specific areas in real time and promptly detects any unauthorized entry or abnormal activities. Whether it is a home, commercial place or industrial facility, a reliable intrusion detection system is needed to ensure the safety of people and property. The intrusion detection system can quickly issue an alarm when a potential security threat occurs so that appropriate measures can be taken to prevent and respond. The night environment poses a huge challenge to intrusion detection. Under low light conditions, traditional detection methods often find it difficult to accurately identify target objects and judge their behavior. The laser-vision linkage intrusion detection method provides new ideas and methods for solving the problem of night detection.

[0003] Laser sensors have high-precision distance measurement capabilities and can accurately determine the position of target objects at night. Even in the dark, the laser beam can be reliably reflected back, providing accurate distance information between the target object and the sensor. Visual sensors can collect image information of the target area and provide rich image features, which helps to more accurately identify and classify target objects. Linking laser and visual sensors can give full play to the advantages of both. At night, when the ambient light intensity is low, laser sensors can provide stable distance information, while visual sensors can obtain clearer images by enhancing sensitivity and other methods. This linkage method not only improves the accuracy of night intrusion detection, but also enhances the adaptability and reliability of the system.

[0004] Currently, traditional laser-vision-linked nighttime intrusion detection systems still have certain drawbacks. Existing laser-vision-linked nighttime intrusion detection systems simply combine lasers and visual sensors, often failing to automatically adjust operating parameters based on changes in the environment and target objects. They are unable to flexibly adapt to actual needs under varying lighting intensities, target object movement speeds, and distances, resulting in unstable detection results. The lack of an effective linkage mechanism prevents the full synergy between lasers and visual sensors, hindering optimal detection results. Therefore, it is necessary to propose a laser-vision-linked nighttime intrusion detection system to address these issues. Summary of the Invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a laser-vision linkage night intrusion detection system, which can automatically adjust the working parameters of the laser and vision sensors according to factors such as the ambient light intensity, the moving speed and distance of the target object to achieve the best detection effect, and by establishing a dynamic linkage model, dynamically adjust the coordination between the two to adapt to different intrusion scenarios.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solutions: a laser-vision linkage nighttime intrusion detection system, comprising a laser sensor, a vision sensor, a data processing module, and a control module. The laser sensor is used to emit a laser beam and receive reflected laser signals to measure the distance and position of the target object. The vision sensor is used to collect image information of the target area. The data processing module processes and analyzes the data collected by the laser and vision sensors and, together with the control module, establishes a dynamic linkage model. The control module automatically adjusts the operating parameters of the laser and vision sensors based on the ambient light intensity, the target object's movement speed, and the distance, dynamically adjusting the coordination between the two.

[0007] The control module includes a light intensity detection unit, a target object speed detection unit, a target object distance detection unit and a parameter adjustment unit. The light intensity detection unit is used to detect the light intensity of the environment, the target object speed detection unit is used to detect the moving speed of the target object, and the target object distance detection unit is used to detect the distance between the target object and the sensor. The parameter adjustment unit, through the coordinated work with the light intensity detection unit, the target object speed detection unit and the target object distance detection unit, realizes the precise adjustment of the sensor parameters to adapt to different environments and target object states;

[0008] The light intensity detection unit in the control module detects the ambient light intensity as follows: , Less than the normal light intensity threshold When , a signal is sent to the parameter adjustment unit, and the parameter adjustment unit adjusts the parameter according to the formula Increase the power of the laser sensor, where is the laser power adjustment coefficient, and according to the formula Enhance the sensitivity of visual sensors, including is the sensitivity adjustment coefficient of the visual sensor, and the target object speed detection unit detects the target object speed as , Greater than the speed threshold When the parameter adjustment unit is based on the formula Increase the sampling frequency of the sensor, where is the sampling frequency adjustment coefficient, and the target object distance detection unit detects the target object distance as , when in different distance intervals, , the parameter adjustment unit is based on the formula Adjust the laser power according to the formula Adjust the visual sensor sensitivity, where and is the adjustment coefficient corresponding to the distance interval.

[0009] Furthermore, the laser sensor is composed of a laser emitting unit, a receiving unit and a signal processing unit, which can emit a laser beam of a specific wavelength and frequency, receive and process the reflected laser signal, and calculate the distance and position information between the target object and the sensor. The laser sensor measures the round-trip time of the laser. , according to the formula Determine the distance between the target object and the sensor, where The speed of light.

[0010] Furthermore, the visual sensor is composed of a lens, an image sensor and an image processing unit, which can collect light in the target area, convert the light signal into an electrical signal to generate image data, and process and analyze it to extract the characteristic information of the target object. The visual sensor assists in judging the light intensity by analyzing the brightness and contrast characteristics of the image.

[0011] Furthermore, the data processing module extracts high-precision distance information of laser data and image features of visual data in real time, and fuses and analyzes these data. The data processing module extracts texture features by calculating the gray level co-occurrence matrix of the image, and defines the gray level of the image as , the image size is , the relative position offset of two pixels is , then the gray-level co-occurrence matrix Indicates that the gray value in the image is The pixel point and the distance to it The gray value at position The probability of simultaneous occurrence of pixels is calculated, and the texture characteristics of the image are described by calculating various statistics of the gray-level co-occurrence matrix.

[0012] Furthermore, the light intensity detection unit uses the parameters of the image sensor to assist in judging the light intensity. The system analyzes the collected image and counts the distribution of brightness values ​​of pixels in the image. If the image is darker as a whole, it means that the light intensity is low, and if the image is brighter, it means that the light intensity is high.

[0013] Furthermore, the target object speed detection unit detects the moving speed of the target object using the Doppler effect and motion detection algorithm;

[0014] The Doppler effect is used to detect speed, and the laser frequency emitted by the laser sensor is defined as , the irradiation speed is When the laser is reflected back from the target object, the received laser frequency Through the Doppler effect formula Calculated by measuring and The speed of the target object is calculated by ,in is the speed of light;

[0015] Use motion detection algorithm to determine the position coordinates of the target object in continuous image frames and The time interval for image acquisition is , then the moving speed of the target object on the image plane is expressed by the formula Calculated.

[0016] Furthermore, the target object distance detection unit uses a laser sensor to measure the distance between the target object and the sensor by measuring the laser round trip time. , and combined with the image analysis technology of the visual sensor, the distance is estimated through stereo vision and monocular vision;

[0017] Stereo vision calculates the distance between the target object and the sensor through the principle of triangulation ,in is the baseline distance between the two vision sensors, is the focal length of the visual sensor, is the disparity of the target object in the two images;

[0018] Monocular vision estimates the distance based on the size of the target object in the image, the known actual size of the target object, and the focal length of the visual sensor. ,in is the actual size of the target object, is the height of the target object in the image, is the focal length of the vision sensor.

[0019] Furthermore, the parameter adjustment unit comprehensively considers the light intensity, target object speed and distance factors, and adjusts the parameters of the laser and visual sensor as a whole. The final laser power is adjusted to , the visual sensor sensitivity is adjusted to , the sampling frequency is adjusted to .

[0020] Furthermore, the data processing module and the control module jointly establish a dynamic linkage model as follows:

[0021] First, the parameter adjustment unit in the control module adjusts the operating parameters of the laser and vision sensor based on environmental data, establishing the basic framework of the linkage model. This enhances the laser power and vision sensor sensitivity in low-light environments, increases the sensor sampling frequency when the target object moves quickly, and adjusts the coordination between the two based on the distance to the target object.

[0022] The data processing module then fuses the high-precision distance information of the laser data with the image features of the visual data, optimizes the linkage model through a multimodal data fusion algorithm, and uses machine learning and artificial intelligence technologies to learn and analyze historical data to continuously improve the performance of the linkage model.

[0023] As the system runs, the environment and target objects are continuously monitored in real time, and the linkage model is dynamically adjusted and improved based on new data.

[0024] Furthermore, the multimodal data fusion algorithm adopts a weighted average method and defines the target object position estimate provided by the laser data as , the target object position estimated by the visual data is , the fused position is estimated to be ,in and is the weight coefficient, which is adjusted according to the reliability and accuracy of laser and visual data. The machine learning algorithm is used for behavior pattern recognition. The support vector machine SVM algorithm is used to define the training sample set as ,in is the eigenvector, is the category label, and the goal of SVM is to find a hyperplane , so that the classification interval is maximized, by solving the optimization problem: ,in is a slack variable, and satisfies , get the optimal and , to achieve the classification and recognition of target object behavior.

[0025] Compared with existing technologies, this laser-vision-linked nighttime intrusion detection system has the following beneficial effects:

[0026] The present invention utilizes the coordination between the various units in the control module to achieve the effect of automatically adjusting sensor parameters according to the environment and the state of the target object, enhances the power of the laser and the sensitivity of the visual sensor in a low-light environment, increases the sampling frequency of the sensor when the target object moves quickly, and adjusts the coordination mode of the laser and visual sensor according to the distance of the target object, thereby improving the accuracy and adaptability of detection. Through the fusion analysis of laser and visual data by the data processing module, and the establishment of a dynamic linkage model together with the control module, the high-precision distance information of the laser data and the rich image features of the visual data are utilized to achieve high-precision detection and recognition of the target object.

[0027] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0029] Figure 1 This is a flow chart of a laser-vision-linked nighttime intrusion detection system;

[0030] Figure 2 This is a structural diagram of a laser-vision linkage nighttime intrusion detection system. DETAILED DESCRIPTION

[0031] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0032] Example 1

[0033] In order to strengthen security protection at night, a large factory installed the laser-vision linkage nighttime intrusion detection system of the present invention.

[0034] Laser sensors and vision sensors are deployed at key locations around the factory, such as around walls and at entrances to important equipment areas. The installation angle of the laser sensors has been carefully adjusted to ensure that they can cover the widest possible range while ensuring measurement accuracy. The vision sensors are installed in positions where they can clearly capture images of the target area, and there is a certain overlap with the coverage area of ​​the laser sensors to facilitate better coordinated detection.

[0035] The laser sensor and vision sensor are connected to the data processing module and control module through a high-speed data transmission line. After the system is started, a comprehensive self-test and initialization are performed. First, each sensor tests its own performance to ensure that the laser emission and reception are normal and the image acquisition of the vision sensor is clear. Then, the data processing module and control module load the preset parameters and algorithms to prepare for subsequent detection work.

[0036] The light intensity detection unit continuously monitors the light intensity around the factory through the light sensor. , the normal light intensity threshold is ,like For example, when the light intensity is low at night, the system starts the light compensation mechanism. The photosensor converts the received light signal into an electrical signal and determines the light intensity value according to a preset conversion formula.

[0037] The target object speed detection unit and the target object distance detection unit start working, using the Doppler effect and motion detection algorithm to detect the moving speed and distance of the possible target object. If a moving target is detected, the target object speed is calculated based on the frequency change of the laser reflection signal and the displacement of the target object in the image. and distance from the factory .

[0038] For the Doppler effect detection speed, the laser frequency emitted by the laser sensor is When the laser is reflected back from the target object, the received laser frequency The Doppler effect formula can be used (in is the speed of light) calculated by measuring and The speed of the target object can be calculated by .

[0039] For motion detection algorithms, determine the position coordinates of the target object in consecutive image frames. and The time interval for image acquisition is , then the moving speed of the target object on the image plane can be expressed by the formula The target object distance detection unit uses a laser sensor to measure the distance and record the time difference between laser emission and reception. , according to the speed of light , the distance between the target object and the sensor ,At the same time, the image analysis technology of the ,vision sensor can also be combined to estimate the distance ,through stereo vision or the size and position relationship of the ,target object in the image.

[0040] Stereo vision calculates the distance between the target object and the sensor through the principle of triangulation ,in is the baseline distance between the two vision sensors, is the focal length of the visual sensor, is the disparity of the target object in the two images.

[0041] Monocular vision estimates the distance based on the size of the target object in the image, the known actual size of the target object, and the focal length of the visual sensor. ,in is the actual size of the target object, is the height of the target object in the image, is the focal length of the vision sensor.

[0042] The laser sensor emits a laser beam, and the laser round trip time , using the formula (in The laser sensor calculates the distance between the target object and the sensor and determines the approximate position of the target object. At the same time, the signal processing unit of the laser sensor analyzes the received laser signal and extracts some characteristic information of the target object, such as reflection intensity. This information can help determine the material and other characteristics of the target object.

[0043] The visual sensor collects image information of the target area, analyzes the brightness, contrast and other features of the image to assist in judging the light intensity, and uses image processing algorithms to extract image features. For example, the gray level co-occurrence matrix of the image is calculated to extract texture features. Assume that the gray level of the image is , the image size is , the relative position offset of two pixels is , then the gray-level co-occurrence matrix Indicates that the gray value in the image is The pixel point and the distance to it The gray value at position The probability of simultaneous occurrence of pixels is calculated, and the texture features of the image are described by calculating various statistics of the gray-level co-occurrence matrix to determine the type of target object. Image recognition technology can also be used to identify whether the target object is a person, a vehicle, or other objects.

[0044] The control module adjusts the sensor parameters according to the environmental data. If the light intensity meets the , the laser power enhancement coefficient is based on the formula Calculate and enhance the power of the laser sensor; the sensitivity enhancement coefficient of the visual sensor is , enhance the sensitivity of visual sensors, where and To adjust the coefficient, it can be set according to the actual environment and needs of the factory.

[0045] When the target object speed Greater than the speed threshold When the sampling frequency increases by , increase the sampling frequency of the sensor, The speed adjustment coefficient ensures that when the target object moves quickly, the system can capture its motion state in a timely and accurate manner.

[0046] According to the distance of the target object Adjust the way the two work together. In a specific range, such as , the laser power is adjusted to , the visual sensor sensitivity is adjusted to , and The distance adjustment coefficient dynamically adjusts the parameters of the laser and visual sensors as the distance to the target object changes to achieve the best detection effect.

[0047] The data processing module fuses the high-precision distance information of the laser data with the image features of the visual data, and adopts the weighted average method. The target object position estimated by the laser data is , the target object position estimated by the visual data is , the fused position is estimated to be ,in , the weight coefficients are adjusted according to the reliability and accuracy of the laser and vision data. For example, if the laser data has high accuracy but may be disturbed in some cases, while the vision data is relatively stable but slightly less accurate, then the values ​​of α and β can be adjusted according to the actual situation to obtain a more accurate position estimate.

[0048] Use machine learning algorithm to identify behavior patterns, adopt support vector machine (SVM) algorithm, and set the training sample set as ,in is the eigenvector, is the category label, and the goal of SVM is to find a hyperplane , so that the classification interval is maximized, by solving the optimization problem: ,in is a slack variable, and satisfies , we can get the optimal and , thereby judging whether the behavior of the target object is abnormal intrusion behavior. For example, by learning from historical data, a normal behavior pattern library is established. When a new target object appears, its features are compared with the patterns in the library. If the difference is large, it is judged as abnormal intrusion behavior.

[0049] The system continuously monitors the factory's surrounding environment and target objects in real time. If the ambient light intensity, target object speed or distance changes, the control module will immediately adjust the sensor parameters and notify the data processing module. The data processing module will dynamically adjust the linkage model based on the new data.

[0050] For example, when the target object suddenly accelerates, the system recalculates the sampling frequency increase coefficient ,in The new target object speed is calculated, and the sensor sampling frequency is adjusted according to the new coefficient. If the target object's behavior is judged to be abnormal intrusion, the system will immediately issue an alarm and notify the factory security personnel to take corresponding measures. The alarm can be issued through various means such as sound and light alarm devices, mobile phone text messages, and monitoring center display screens to ensure that security personnel can receive and respond in time.

[0051] In summary: The laser-vision linkage detection system can detect any potential intrusion in a timely and accurate manner, whether it is a person climbing a wall or a vehicle entering illegally, it can be detected and an alarm can be issued in the first time. The sensor parameters are automatically adjusted according to factors such as ambient light intensity, target object speed and distance to ensure that the system can maintain good performance in various complex situations. Whether in a low-light environment or facing fast-moving target objects, efficient detection can be achieved. The use of machine learning algorithms for behavioral pattern recognition can accurately distinguish between normal activities and abnormal intrusion behaviors, avoiding unnecessary alarms and missed reports. The system's real-time monitoring and alarm functions enable security personnel to respond quickly, reduce patrol workload, improve the efficiency of safety management, ensure that the factory can maintain a safe and stable environment at night, and avoid production interruptions and property losses caused by intrusion incidents.

[0052] Example 2

[0053] In an important highway tunnel, the laser-vision linkage nighttime intrusion detection system of the present invention is installed to ensure driving safety and prevent illegal intrusion.

[0054] Laser sensors and vision sensors are installed at the entrance, exit and inside the tunnel at certain intervals. The laser sensor is installed high on the tunnel wall to ensure that it can cover the entire tunnel cross-section and will not be interfered with by passing vehicles. The vision sensor is installed in a suitable position so that it can clearly capture vehicles and other target objects in the tunnel.

[0055] After the system is started, it performs self-test and initialization. Each sensor checks its own working status, and the data processing module and control module load preset parameters and algorithms to prepare for subsequent monitoring work.

[0056] The light intensity detection unit continuously monitors the light intensity in the tunnel. Since the light in the tunnel is usually provided by the lighting system, the light intensity is relatively stable. However, in some cases, such as lighting system failure or special weather conditions, the light intensity may change. Suppose the current light intensity in the tunnel is , the normal light intensity threshold is ,like , the system automatically adjusts the sensor parameters to adapt to low light conditions.

[0057] The target object speed detection unit and the target object distance detection unit start working, using the Doppler effect and motion detection algorithm to detect the moving speed and distance of the vehicle in the tunnel. For the Doppler effect detection speed, the laser frequency emitted by the laser sensor is set to , when the irradiation speed is When the laser beam is reflected back onto the vehicle, the received laser frequency The Doppler effect formula can be used (in is the speed of light) and is calculated by measuring and The difference between the two values ​​can be used to calculate the vehicle speed. .

[0058] The target object distance detection unit uses a laser sensor to measure the distance and record the time difference between laser emission and reception. , according to the speed of light , the distance between the vehicle and the sensor ,At the same time, the image analysis technology of the ,visual sensor can be combined to estimate the distance ,according to the size and position relationship of the vehicle in the ,image.

[0059] The laser sensor emits a laser beam, and the laser round trip time , using the formula (in The laser sensor can also detect information such as the height and width of the vehicle, providing more data for subsequent analysis.

[0060] Visual sensors collect image information in the tunnel, analyze image features such as brightness and contrast to assist in determining light intensity, and use image processing algorithms to extract image features. For example, image recognition technology can be used to identify vehicle type, color, license plate number, and other information. The vehicle's trajectory and behavior can also be analyzed to determine whether there are any abnormal conditions, such as speeding or illegal lane changes.

[0061] The control module adjusts the sensor parameters according to the environmental data. If the light intensity meets the , the laser power enhancement coefficient is based on the formula Calculate and enhance the power of the laser sensor; the sensitivity enhancement coefficient of the visual sensor is , enhance the sensitivity of visual sensors, where and To adjust the coefficient, it can be set according to the actual lighting conditions in the tunnel.

[0062] When the vehicle speed Greater than the speed threshold When the sampling frequency increases by , increase the sampling frequency of the sensor, The speed adjustment coefficient ensures that the system can capture the vehicle's motion status in a timely and accurate manner when the vehicle is traveling at high speed.

[0063] According to vehicle distance Adjust the coordination between the two. If the vehicle is far away, the laser sensor will be used for detection first to obtain the vehicle's accurate location information. When the vehicle is close, the image information of the visual sensor will be combined to improve the accuracy of vehicle recognition.

[0064] The data processing module fuses the high-precision distance information of the laser data with the image features of the visual data, and adopts the weighted average method. The vehicle position estimate provided by the laser data is , the vehicle position estimate provided by the visual data is , the fused position is estimated to be ,in , the weight coefficients are adjusted according to the reliability and accuracy of the laser and vision data. For example, if the laser data is more accurate when measuring at long distances but may be interfered with at close distances, while the vision data can provide more detailed information at close distances, then the values ​​of α and β can be adjusted according to the distance of the vehicle to obtain a more accurate position estimate.

[0065] Use machine learning algorithm to identify behavior patterns, adopt support vector machine (SVM) algorithm, and set the training sample set as in is the eigenvector, It is a category label. By learning from historical data, a library of normal vehicle driving patterns is established. When a new vehicle appears, its characteristics are compared with the patterns in the library. If the difference is large, it is judged as abnormal behavior, such as speeding, illegal lane changing, etc.

[0066] The system continuously monitors the environment and vehicles in the tunnel in real time. If the ambient light intensity, vehicle speed or distance changes, the control module will immediately adjust the sensor parameters and notify the data processing module. The data processing module will dynamically adjust the linkage model based on the new data.

[0067] For example, when a vehicle is detected speeding, the system immediately issues an alarm to notify the tunnel management department to take appropriate measures. The alarm can be issued through the sound and light alarm device in the tunnel, the monitoring center display screen, etc. At the same time, the information of the violating vehicle can be transmitted to the traffic management department for subsequent processing.

[0068] In summary: By real-time monitoring of vehicles and the environment in the tunnel, potential safety hazards such as speeding, illegal lane changes, and lighting system failures can be discovered in a timely manner, providing drivers with a safer driving environment. Sensor parameters are automatically adjusted according to factors such as light intensity, vehicle speed, and distance in the tunnel to ensure that the system can maintain good performance in various complex situations. Whether in low-light conditions or facing high-speed vehicles, efficient monitoring can be achieved. The system's real-time monitoring and alarm functions enable tunnel management departments to respond quickly and deal with various abnormal situations in a timely manner, reducing the probability of accidents, improving tunnel traffic efficiency, ensuring that highway tunnels can maintain safe and stable operation at night, and avoiding traffic interruptions and property losses caused by safety accidents.

[0069] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and scope of the same elements of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A laser-vision linkage nighttime intrusion detection system, characterized in that: The system includes a laser sensor, a visual sensor, a data processing module and a control module. The laser sensor is used to emit a laser beam and receive the reflected laser signal to measure the distance and position information of the target object. The visual sensor is used to collect image information of the target area. The data processing module processes and analyzes the data collected by the laser and visual sensors, and establishes a dynamic linkage model with the control module. The control module automatically adjusts the working parameters of the laser and visual sensors according to the light intensity of the environment, the moving speed and distance of the target object, and dynamically adjusts the coordination between the two. The control module includes a light intensity detection unit, a target object speed detection unit, a target object distance detection unit and a parameter adjustment unit, wherein the light intensity detection unit is used to detect the light intensity of the environment, the target object speed detection unit is used to detect the moving speed of the target object, and the target object distance detection unit is used to detect the distance between the target object and the sensor. The parameter adjustment unit realizes precise adjustment of sensor parameters to adapt to different environments and target object states by working in coordination with the light intensity detection unit, the target object speed detection unit and the target object distance detection unit; The light intensity detection unit in the control module detects the ambient light intensity as follows: , Less than normal light intensity threshold When , a signal is sent to the parameter adjustment unit, and the parameter adjustment unit adjusts the parameter according to the formula Increase the power of the laser sensor, where is the laser power adjustment coefficient, and according to the formula Enhance the light sensitivity of the visual sensor, including is the sensitivity adjustment coefficient of the visual sensor, and the target object speed detection unit detects the target object speed as , Greater than speed threshold When the parameter adjustment unit is based on the formula Increase the sampling frequency of the sensor, where is the sampling frequency adjustment coefficient, and the target object distance detection unit detects the target object distance as , when in different distance intervals, , the parameter adjustment unit is based on the formula Adjust the laser power according to the formula Adjust the visual sensor sensitivity, where and is the adjustment coefficient corresponding to the distance interval.

2. A laser-vision-linked nighttime intrusion detection system according to claim 1, characterized in that: The laser sensor consists of a laser emitting unit, a receiving unit and a signal processing unit. It can emit a laser beam of a specific wavelength and frequency, receive and process the reflected laser signal, and calculate the distance and position information between the target object and the sensor. The laser sensor measures the round-trip time of the laser. , according to the formula Determine the distance of the target object from the sensor, where The speed of light.

3. The laser-vision linkage nighttime intrusion detection system according to claim 1 is characterized in that: The visual sensor is composed of a lens, an image sensor and an image processing unit. It can collect light in the target area, convert light signals into electrical signals to generate image data, and process and analyze them to extract feature information of the target object. The visual sensor assists in judging light intensity by analyzing the brightness and contrast characteristics of the image.

4. The laser-vision linkage nighttime intrusion detection system according to claim 1 is characterized in that: The data processing module extracts high-precision distance information of laser data and image features of visual data in real time, and fuses and analyzes these data. The data processing module extracts texture features by calculating the grayscale co-occurrence matrix of the image, and defines the grayscale of the image as , the image size is , the relative position offset of two pixels is , then the gray-level co-occurrence matrix Indicates that the gray value in the image is The pixel point and its distance The gray value at position The probability of pixels appearing at the same time is calculated, and the texture characteristics of the image are described by calculating various statistics of the gray-level co-occurrence matrix.

5. The laser-vision linkage nighttime intrusion detection system according to claim 1 is characterized in that: The light intensity detection unit uses the parameters of the image sensor to assist in determining the light intensity. The system analyzes the collected image and counts the distribution of the brightness values ​​of the pixels in the image. If the image is darker as a whole, it means that the light intensity is low, and if the image is brighter, it means that the light intensity is high.

6. The laser-vision linkage nighttime intrusion detection system according to claim 1, characterized in that: The target object speed detection unit detects the moving speed of the target object using the Doppler effect and a motion detection algorithm; Using the Doppler effect to detect speed, the laser frequency emitted by the laser sensor is defined as , the irradiation speed is When the laser is reflected back from the target object, the received laser frequency The Doppler effect formula Calculated by measuring and The difference between ,in is the speed of light; Use motion detection algorithm to determine the position coordinates of the target object in continuous image frames and The time interval for image acquisition is , then the moving speed of the target object on the image plane is expressed by the formula Calculated.

7. The laser-vision-linked nighttime intrusion detection system according to claim 1 is characterized in that: The target object distance detection unit uses a laser sensor to measure the distance between the target object and the sensor by measuring the laser round trip time. , and combined with the image analysis technology of the visual sensor, the distance is estimated through stereo vision and monocular vision; Stereo vision calculates the distance between the target object and the sensor through the principle of triangulation ,in is the baseline distance between the two vision sensors, is the focal length of the visual sensor, is the disparity of the target object in the two images; Monocular vision estimates the distance based on the size of the target object in the image, the known actual size of the target object, and the focal length of the visual sensor. ,in is the actual size of the target object, is the height of the target object in the image, is the focal length of the visual sensor.

8. The laser-vision linkage nighttime intrusion detection system according to claim 1 is characterized in that: The parameter adjustment unit comprehensively considers the light intensity, target object speed and distance factors, and adjusts the parameters of the laser and visual sensor as a whole. The final laser power is adjusted to , the visual sensor sensitivity is adjusted to , the sampling frequency is adjusted to .

9. The laser-vision linkage nighttime intrusion detection system according to claim 1, characterized in that: The process of establishing a dynamic linkage model jointly established by the data processing module and the control module is as follows: First, the parameter adjustment unit in the control module adjusts the working parameters of the laser and visual sensor according to the environmental data, establishes the basic framework of the linkage model, enhances the laser power and the sensitivity of the visual sensor in a weak light environment, increases the sensor sampling frequency when the target object moves quickly, and adjusts the coordination between the two according to the distance of the target object; Then, the data processing module fuses the high-precision distance information of the laser data and the image features of the visual data, optimizes the linkage model through a multimodal data fusion algorithm, and uses machine learning and artificial intelligence technology to learn and analyze historical data to continuously improve the performance of the linkage model; As the system runs, the environment and target objects are continuously monitored in real time, and the linkage model is dynamically adjusted and improved based on new data.

10. The laser-vision linkage nighttime intrusion detection system according to claim 9, characterized in that: The multimodal data fusion algorithm adopts a weighted average method and defines the target object position estimate provided by the laser data as , the target object position estimated by the visual data is , the fused position is estimated to be ,in and is the weight coefficient, which is adjusted according to the reliability and accuracy of laser and visual data. The machine learning algorithm is used for behavior pattern recognition. The support vector machine SVM algorithm is used to define the training sample set as ,in is the eigenvector, is the category label, and the goal of SVM is to find a hyperplane , so that the classification interval is maximized, by solving the optimization problem: ,in is a slack variable, and satisfies , get the best and , to achieve the classification and recognition of target object behavior.

Images