A timing photographing error compensation system and method for a uniform motion scene

Abstract

The application discloses a timing photographing error compensation system for a uniform motion scene, comprising a motion sensing sensor, a central processing unit and a camera. The motion sensing sensor collects the initial position and time of a photographed object and generates a digital signal, which is transmitted to the central processing unit through a GPIO interface; the central processing unit calculates a delay photographing time parameter and runs an error compensation algorithm to optimize the position of the photographed object in an image; and the camera photographs an image under the triggering of the central processing unit. The timing photographing error compensation system effectively improves the photographing precision and image quality by dynamically adjusting the delay photographing time parameter, and is suitable for precise photographing of a uniform motion object.

Description

Technical Field

[0001] This application relates to the field of image capture and error compensation technology, and in particular to a timed photography error compensation system and method for uniform motion scenes. Background Technology

[0002] In industrial production, logistics sorting, and quality inspection, there are numerous scenarios involving uniform motion. In these scenarios, capturing images of moving objects at regular intervals is a critical requirement. However, existing technologies typically achieve this by continuously and rapidly taking repeated shots or by sensors triggering immediate capture, but these methods have significant drawbacks. While continuous rapid shooting can capture moving objects, it generates a large number of invalid photos, resulting in a low effective photo rate and an unreliable hit rate for single shots, while also wasting storage resources and the number of shots.

[0003] Furthermore, while sensor-triggered photography improves the hit rate of a single shot, it places extremely high demands on the sensor's installation location and system response speed, increasing the complexity of equipment installation and maintenance. When the object's speed changes, the object's position in the photograph will deviate significantly, failing to consistently appear in the center of the camera lens, severely impacting image quality. Summary of the Invention

[0004] To address existing problems, this application proposes a system and method for timed image capture error compensation in uniform motion scenarios. By accurately calculating and fine-tuning the delay time, it ensures that the object being photographed consistently appears at the center of the camera's 130mm lens, significantly improving image quality and single-shot hit rate. Furthermore, the sensor position can be flexibly adjusted according to environmental conditions, enhancing the device's applicability and adaptability while reducing installation and maintenance difficulty, providing a more efficient and stable solution for image detection in industrial production, logistics sorting, and other fields.

[0005] This application provides a timing-based photo-taking error compensation system for uniform motion scenes, including: a motion sensing sensor for acquiring the initial position and initial time of the object being photographed and generating corresponding signals; wherein, the motion sensing sensor converts the acquired analog signals of the object being photographed into digital signals, the digital signals are transmitted based on the GPIO interface of the processor inside the motion sensing sensor, and the digital signals are transmitted to the central processing unit by an interrupt response mechanism. The central processing unit (CPU) is used to receive the digital signal and calculate the time-lapse shooting parameters, wherein the time-lapse shooting parameters are calculated based on the distance and relative speed between the subject and the camera; the CPU is also used to run an error compensation algorithm to dynamically adjust the time-lapse shooting parameters based on multiple shooting results in order to optimize the position of the subject in the image. A camera is used to capture an image of the object being photographed when triggered by the central processing unit; wherein, the error compensation algorithm includes an association model based on time deviation and position deviation, and uses a sliding window weighted average method to dynamically correct the time-lapse shooting parameters.

[0006] Furthermore, the motion sensing sensor includes a position acquisition sensor and a time acquisition sensor, used to acquire the initial position and initial time of the object being photographed.

[0007] Furthermore, the motion sensing sensor also includes an analog-to-digital converter, which converts the analog signals acquired by the position acquisition sensor and the time acquisition sensor into digital signals.

[0008] Furthermore, the digital signal is transmitted to the central processing unit through the GPIO interface, which supports multiple communication protocols and triggers the central processing unit to read data through an interrupt response mechanism.

[0009] Furthermore, the central processing unit is specifically used to: receive the digital signal and calculate the time delay shooting parameters; run an error compensation algorithm to adjust the time delay shooting parameters based on multiple shooting results, and optimize the position of the photographed object in the image.

[0010] Furthermore, the central processing unit includes: an error acquisition module for acquiring the positional and temporal deviations of the photographed object; an error modeling module for establishing a correlation model between the positional and temporal deviations; and a parameter correction module for dynamically adjusting the time-lapse shooting parameters.

[0011] Furthermore, the camera captures images of the subject under the trigger of the central processor. The camera 130 supports multiple shooting modes, including single shooting, continuous shooting, and timed shooting.

[0012] Furthermore, the camera's triggering mechanism works in conjunction with the central processing unit's error compensation algorithm to ensure the accuracy and consistency of the captured images.

[0013] This application provides a method for compensating for timing-based photography errors in uniform motion scenes, characterized by the following steps: The initial position and initial time of the object being photographed are collected by the motion sensing sensor, a digital signal is generated and transmitted to the central processing unit via the GPIO interface; The central processing unit receives the digital signal and calculates the time-lapse shooting parameters based on the distance and relative speed between the object being photographed and the camera. The central processing unit runs an error compensation algorithm to adjust the time delay shooting parameter based on multiple shooting results, thereby optimizing the position of the photographed object in the image. The camera captures an image of the object under the trigger of the central processing unit; the error compensation algorithm includes an association model based on time deviation and position deviation, and uses a sliding window weighted average method to dynamically correct the time delay shooting parameter.

[0014] Furthermore, the method includes: the central processing unit determining an initial optimal shooting time point based on the time-lapse shooting time parameter, and triggering the camera to perform a shooting operation at the initial optimal shooting time point; the central processing unit dynamically adjusting the time-lapse shooting time parameter based on the positional and temporal deviations of multiple shooting results, and the central processing unit determining a corrected optimal shooting time point based on the adjusted time-lapse shooting time parameter, and triggering the camera to perform a shooting operation at the corrected optimal shooting time point. Attached Figure Description

[0015] Figure 1 This is a conceptual block diagram of a timing-based photographic error compensation system for uniform motion scenes, based on some examples of this disclosure.

[0016] Figure 2 This is a conceptual block diagram of a motion-sensing sensor processing an initial signal of a photographed object, based on some examples of this disclosure.

[0017] Figure 3 This is a conceptual block diagram of a central processing unit (CPU) processing and running an error compensation algorithm, based on some examples of this disclosure.

[0018] Figure 4 This is a conceptual block diagram of a camera options interface for a central processing unit, based on some examples of this disclosure.

[0019] Figure 5 This is a conceptual block diagram illustrating the initial relationship between a timing-based photographing error compensation system for uniform motion scenes and the photographed object, based on some examples of this disclosure.

[0020] Figure 6 This is a conceptual block diagram of an object being photographed arriving at a shooting position, based on some examples of this disclosure.

[0021] Figure 7 This is a flowchart illustrating an exemplary process of a timed photography error compensation method for uniform motion scenes, based on some examples of this application.

[0022] Figure 8 This is a flowchart illustrating an exemplary process for the acquisition and transmission of motion sensing sensor data according to some examples of this application.

[0023] Figure 9 This is a flowchart illustrating an exemplary process of a central processing unit running an error compensation algorithm according to some examples of this application.

[0024] Figure 10 This is a flowchart illustrating an exemplary process of a camera-triggered shooting procedure according to some examples of this application. Detailed Implementation

[0025] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are also within the scope of protection of this application.

[0026] The terminology used in this specification is that which is currently widely used in the art in consideration of the functionality of this application; however, these terms may vary depending on the intent of a person skilled in the art, precedent, or new technology in the art. Furthermore, specific terms may be chosen, and in such cases, their detailed meanings will be described in the detailed description of this application. Therefore, the terminology used in this specification should not be construed as simple names, but rather based on the meaning of the terms and the overall description of this application.

[0027] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0028] This application uses flowcharts to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously, as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0029] Furthermore, the application architecture diagrams in this application are for the purpose of more clearly illustrating the technical solutions in this application and do not constitute a limitation on the technical solutions provided in this application. Of course, the technical solutions provided in this application are also applicable to similar problems for other application architectures and business applications.

[0030] The following examples or embodiments illustrate a non-limiting method for timed photography error compensation in uniform motion scenes according to at least one embodiment of this application. As described below, different features in these specific examples or embodiments can be combined with each other without conflict to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of this application.

[0031] This invention discloses a timed image capture error compensation system for uniform motion scenes, comprising a motion sensing sensor 110, a central processing unit 120, and a camera 130. The motion sensing sensor 110 acquires the initial position and time of the object being photographed and generates a digital signal, which is transmitted to the central processing unit 120 via a GPIO interface 114 built into the motion sensing sensor 110. The central processing unit 120 calculates the time delay shooting parameters and runs an error compensation algorithm 1200 to optimize the position of the object being photographed in the image. The camera 130 captures the image upon triggering by the central processing unit 120. This timed image capture error compensation system effectively improves shooting accuracy and image quality by dynamically adjusting the time delay shooting parameters, and is suitable for precise shooting of objects moving at a uniform speed.

[0032] The technical solution proposed in this application aims to address the problems of low hit rate, poor image quality, and serious resource waste in timed photography under uniform motion scenarios in existing technologies. By introducing a collaborative working mechanism between the motion sensing sensor 110, the central processing unit 120, and the camera 130, combined with an error compensation algorithm 1200, dynamic adjustment of the shooting timing is achieved, thereby ensuring the stable and clear position of the photographed object in the image and improving the accuracy and efficiency of image capture. Furthermore, the system supports multiple shooting modes, possesses good environmental adaptability and scalability, and is suitable for various application scenarios such as industrial production, logistics sorting, and traffic monitoring.

[0033] The terminology used in this specification is defined based on common understanding in the art, and some terms may be adjusted according to actual application scenarios. For example, "uniform motion" not only refers to constant speed motion under ideal conditions, but also includes approximately uniform motion with small speed changes over a short period of time; "error compensation" not only refers to the correction of time deviations, but also includes the comprehensive processing of multi-dimensional errors such as positional deviations and angular deviations. Therefore, the terms used in this specification should not be understood as simple names, but should be understood in conjunction with their specific meaning in this application.

[0034] The system modules described in this application embodiment, such as the motion sensing sensor 110, the central processing unit 120, and the camera 130, can be combined or replaced according to actual needs. For example, the motion sensing sensor 110 can be composed of a single sensor or multiple sensors fused together; the central processing unit 120 can be an embedded chip or a high-performance computing platform; the camera 130 can be an industrial camera 130, a high-speed camera 130, or a regular digital camera 130, the specific selection depending on the application scenario and performance requirements. Therefore, the modules described are only illustrative examples and should not constitute a limitation on the technical solution of this application.

[0035] This application uses flowcharts to illustrate the operations performed by the system. It should be understood that the order of operations described in the flowchart is not fixed and can be adjusted according to actual needs. For example, some steps can be executed in parallel, while others can be omitted or added. Furthermore, the logic and steps described in the flowchart can also be implemented using software, hardware, or firmware, and the specific implementation method can be flexibly chosen based on system architecture and performance requirements.

[0036] The technical solution of this application is applicable not only to image capture by a single device, but also to image acquisition systems with multiple devices working collaboratively. For example, in an industrial production line, multiple shooting units can work together to achieve simultaneous shooting of multiple workstations; in a logistics sorting system, multiple cameras 130 can collaboratively identify package information, improving recognition efficiency and accuracy. Therefore, the technical solution of this application has good scalability and applicability.

[0037] The technical solutions of this application can be implemented in whole or in part through software, hardware, or firmware. When implemented in software, they can be implemented in whole or in part as a computer program product. A computer program product includes one or more computer instructions that can be loaded and executed on a computer to implement the processes or functions described in this application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0038] This application provides a timed photography error compensation system and method for uniform motion scenes, and its implementation is further described below.

[0039] like Figure 1As shown, the timed shooting error compensation method for this uniform motion scene achieves accurate capture of moving objects through the coordinated operation of the motion sensing sensor 110, the central processing unit 120, and the camera 130. The motion sensing sensor 110 monitors the motion state of the object being photographed in real time. In some optional embodiments, it further includes at least one position acquisition sensor 111 and at least one time acquisition sensor 112. When the object enters the detection area, the motion sensing sensor 110 synchronously triggers the position acquisition sensor 111 and the time acquisition sensor 112 to start. The position acquisition sensor 111 calculates the initial position coordinates based on the object's characteristics (such as size and shape) using a visual algorithm or physical signals (such as infrared reflected waves). The time acquisition sensor 112 records the initial time t0 of the object entering the detection area through an internal timer, with an accuracy of up to microseconds. After receiving the sensor data from the position acquisition sensor 111 and the time acquisition sensor 112 through the GPIO interface 114, the central processing unit 120 immediately calls the distance-velocity model to calculate the time-delay shooting parameter B and sends parameter B to the trigger control module of the camera 130. For example, in an industrial conveyor belt scenario, after the motion sensing sensor 110 detects a product entering the detection area, the central processing unit 120 automatically calculates the B value based on the conveyor belt speed, triggering the camera 130 to capture an image when the product reaches the shooting position. The B value is calculated using the formula B = S / v, where S is the physical distance between the sensor and the camera 130, and v is the uniform speed of the object being photographed. The motion sensing sensor 110 and the central processing unit 120 communicate via an interrupt response mechanism to achieve a nanosecond or higher level of response, ensuring real-time data acquisition and processing.

[0040] like Figure 2The diagram further reveals the signal acquisition and processing mechanism of the motion sensing sensor 110. The motion sensing sensor 110 further includes a position acquisition sensor 111, a time acquisition sensor 112, and an analog-to-digital converter 113. The position acquisition sensor 111 selects different technical solutions according to the scene requirements: in industrial scenes with stable lighting, a visual sensor is used to capture the object's outline, for example, using a CMOS image sensor or CCD image sensor combined with an edge detection algorithm to extract the object's outline, and then calculating the relative distance between the object and the camera 130 through a feature matching algorithm; in low-light or high-speed scenes, it switches to a lidar sensor, for example, measuring distance using the TOF (Time of Flight) principle, directly measuring distance through the time difference between transmitting and receiving laser pulses. The time acquisition sensor 112 has a built-in high-precision crystal oscillator, automatically recording a timestamp t0 and binding it to the position data when an object triggers a photoelectric switch or pressure sensor. The original analog signal is sampled with 16-bit precision by the analog-to-digital converter 113, converted into a digital signal, and then transmitted in full-duplex mode through the SPI protocol of the GPIO interface 114. When transmitting large amounts of data (such as raw image data from a vision sensor), the GPIO interface enables multi-channel parallel transmission, sending position coordinates and timestamps synchronously through separate channels to avoid data congestion. Preferably, the analog-to-digital converter is a 16-bit ADC chip with a sampling frequency of at least 10kHz to ensure high-precision synchronous acquisition of position and time signals.

[0041] like Figure 3 The diagram illustrates the implementation process of the error compensation algorithm 1200 of the central processing unit 120. After receiving the digital signal, the central processing unit 120 first parses the initial position coordinates and the timestamp t0. Based on the preset distance S between the sensor and the camera 130 (e.g., the physical distance between two workstations on a conveyor belt), combined with the object's motion speed V (calculated by the difference between two adjacent position data points), it generates initial delay parameters according to the formula B = S / v. After the first shot, the system compares the pixel offset Δx between the theoretical imaging position and the actual imaging center point, and records the actual trigger time t1. The central processing unit 120 further includes an error acquisition module 121 and a parameter correction module 123. The error acquisition module 121 converts Δx into a time deviation Δt according to a proportional relationship (Δx = kV(t1-t0-B)), where k is the pixel-to-time conversion coefficient. The parameter correction module 123 employs a sliding window weighted average method, taking the Δt data from the most recent 5 shots, removing outliers, calculating the average deviation, and updating the parameters according to B' = B + α * Δt (α is the attenuation factor) to prevent over-correction. The error compensation algorithm 1200 also employs a sliding window weighted average method, taking the Δt data from the most recent 5 shots, removing outliers, calculating the average deviation, and updating the parameters according to formula B' = B + α * Δt (α is the attenuation factor) to prevent over-correction. 'The parameter is updated by setting B + α·Δt, where α is a decay factor ranging from 0.1 to 0.5 to prevent over-correction. When the correction amount is less than the set threshold for three consecutive times (e.g., Δt < 0.1 ms), the system automatically stops iteration and locks the current parameter to improve system stability.

[0042] like Figure 4 The diagram illustrates the configuration of the system's interactive interface. In the camera 130 control interface, the operator can select the trigger mode via a drop-down menu: when "External Continuous Trigger" is selected, the interface automatically activates the multi-frame shooting settings, allowing the setting of the number of consecutive shots (1-10) and frame interval (10-1000ms); when "TCP Trigger" is selected, the IP address and port number of the target camera 130 must be entered, and the heartbeat detection frequency must be configured. The exposure parameter setting area provides "Auto Exposure" and "Manual Exposure" switching options. In manual mode, the gain (0-30dB) and exposure time (1μS-1S) can be adjusted by dragging the sliders. The input box for the trigger delay parameter B supports three modification methods: direct keyboard input, clicking the ± button to adjust in 1ms increments, or dragging the progress bar for coarse adjustment. After parameter modification, the interface displays a real-time schematic diagram of the theoretical trigger position, with the center area of ​​the camera 130's field of view marked with a green line. The camera 130 control interface supports real-time parameter adjustment and feedback display. The operator can intuitively view the deviation between the theoretical trigger position and the actual imaging position through the graphical interface and make manual fine adjustments.

[0043] like Figure 5 The diagram illustrates the relationship between the camera 130's triggering mechanism and the object's motion. When the object being photographed passes the motion sensing sensor 110 at a constant speed V, the central processing unit 120 starts a countdown at time t0 and sends a trigger pulse to the camera 130 at time T = t0 + B, when the countdown ends. Upon receiving the pulse, the camera 130 first completes autofocus and exposure preparation (taking approximately 2 ms), and then executes the shutter action. For high-speed moving objects (e.g., speed > 60 km / h), the system employs a pre-trigger buffering mechanism: the camera 130 continuously buffers the image data from the most recent 0.5 seconds. When a trigger signal is received, it automatically extracts images from the 3 frames before to the 2 frames after the trigger moment for the central processing unit 120 to select the optimal frame. For example, in a logistics sorting scenario, when a package experiences speed fluctuations due to vibration, the system dynamically adjusts the B' value to automatically compensate for speed changes at the trigger moment, ensuring that the package's barcode area remains within the image recognition area. The pre-trigger caching mechanism can be implemented using an FPGA or other storage methods such as memory. The camera 130 continuously caches the image data of the most recent 0.5 seconds. When a trigger signal is received, it automatically extracts the images from the three frames before the trigger time to the two frames after the trigger time, allowing the central processing unit 120 to select the optimal frame. This mechanism is particularly suitable for high-speed, uniform motion scenarios such as logistics sorting and traffic monitoring, effectively improving the image capture success rate.

[0044] like Figure 6 The diagram illustrates the effect of the optimized triggering mechanism: after N iterations, the delay parameter gradually converges from the initial value B0 to a stable value Bn. In an industrial inspection case, when the conveyor belt speed is 1.2 m / s, the initial product image position deviates from the center by 35 pixels. After 5 corrections, the deviation is reduced to within ±3 pixels. During parameter correction, the central processing unit 120 records the Δt and B value change curves for each iteration. When the correction amount is less than the set threshold for three consecutive times (e.g., Δt < 0.1 ms), the iteration automatically stops and the current parameters are locked. For sudden speed changes (e.g., the conveyor belt accelerates by 10%), the system restarts the parameter correction process within 0.5 seconds by monitoring the data change rate of adjacent position sensors, ensuring rapid recovery of shooting accuracy at the new steady-state speed. In the industrial inspection case, when the conveyor belt speed is 1.2 m / s, the initial product image position deviates from the center by 35 pixels. After 5 corrections, the deviation is reduced to within ±3 pixels, and the system response time is less than 50 ms. For sudden speed changes (such as a 10% acceleration of the conveyor belt), the system monitors the rate of change of data from sensors at adjacent positions and restarts the parameter correction process within 0.5 seconds to ensure that shooting accuracy is quickly restored at the new steady-state speed.

[0045] Meanwhile, this application further provides a deep learning-based image recognition-assisted localization scheme for a timed shooting error compensation system. In some industrial scenarios, the features of the photographed object are not obvious or are occluded, making it difficult for traditional geometric model-based localization methods to meet high-precision requirements. This embodiment proposes a deep learning-based image recognition-assisted localization scheme to improve the system's recognition and localization capabilities in complex feature scenarios. The camera 130 supports image preprocessing and feature extraction functions, and its image acquisition module uses a high-resolution CMOS image sensor; the central processing unit 120 combines the recognition results and sensor data to calculate the time delay shooting parameter B; the central processing unit 120 further has a deep learning recognition module, which can optionally deploy a lightweight CNN model for recognizing object features. In this embodiment, the high-resolution CMOS image sensor used in the image acquisition module has good low-light performance and dynamic range, enabling it to capture clear object images. The deep learning recognition module deploys a lightweight convolutional neural network (CNN) model, such as MobileNet or YOLO-Lite, for image feature extraction and classification, recognizing object types and key feature points. The central processing unit 120 combines the deep learning recognition results with the position and time data collected by the sensor to calculate a more accurate time-lapse shooting parameter B. The camera 130, under the control of the central processing unit 120, performs the shooting operation to ensure that the object in the image is in the optimal position. The technical advantages of this embodiment are: (1) Deep learning-assisted localization improves the system's recognition capability in complex feature scenes; (2) Lightweight model deployment is suitable for embedded devices, reducing system power consumption; (3) Image preprocessing improves image quality and enhances recognition accuracy; (4) Multimodal fusion combines image recognition and sensor data to improve the overall accuracy of the system. This embodiment is particularly suitable for scenarios such as industrial quality inspection, robot vision, and intelligent warehousing.

[0046] In some embodiments, the timed image capture error compensation system further features a distributed error compensation architecture based on edge computing. In large-scale industrial production lines or logistics sorting systems, centralized processing architectures for timed image capture error compensation systems suffer from large response latency and high data transmission pressure. Therefore, this embodiment proposes a distributed error compensation architecture based on edge computing to improve the system's response speed and stability in large-scale deployment scenarios. The camera 130 is a camera system consisting of multiple cameras 130 working collaboratively, responsible for local data processing and error compensation. The central processing unit 120 further includes a central coordination module 124 responsible for global scheduling and parameter synchronization. Simultaneously, the central processing unit 120 further supports distributed image capture edge computing nodes deployed at the shooting station of each camera 130. Furthermore, the timed image capture error compensation system includes a communication module 140 to support multiple communication protocols such as Ethernet and CAN bus. In some optional embodiments, each edge computing node independently collects data from the camera 130 and motion sensing sensor 110 and executes the error compensation algorithm 1200, reducing dependence on the central processing unit 120 and improving system response speed. The central coordination module 124 periodically synchronizes the parameters of each node to ensure global consistency and avoid system failure due to local errors. Meanwhile, the camera 130 can be composed of a cluster of multiple cameras 130, supporting distributed shooting and improving image acquisition efficiency and coverage. The communication module 140 supports multiple communication protocols, such as Ethernet, CAN bus, and RS485, ensuring the stability and real-time performance of data transmission. The technical advantages of this embodiment are: (1) the edge computing architecture reduces the load on the central processing unit and improves system response speed; (2) distributed deployment is suitable for large-scale industrial scenarios, improving system scalability; (3) the parameter synchronization mechanism ensures the consistency of error compensation parameters between nodes; and (4) multi-camera collaboration improves image acquisition efficiency and coverage, particularly suitable for scenarios such as intelligent manufacturing, large logistics centers, and multi-station inspection systems.

[0047] Accordingly, Figure 7 This is a flowchart illustrating an exemplary process of a timed photography error compensation method for uniform motion scenes, based on some examples of this application.

[0048] exist Figure 7In the example, the motion sensing sensor 110, the central processing unit 120, and the camera 130 work together to achieve timed shooting error compensation for uniform motion scenes. Specifically, the motion sensing sensor 110 is used to collect the initial position and initial time of the object being photographed in real time and transmit the data to the central processing unit 120 (S110); the central processing unit 120 calculates the delayed shooting time parameters based on the received data and dynamically adjusts the parameters through the error compensation algorithm 1200 to optimize the shooting timing (S120); the camera 130, based on the control of the central processing unit 120, triggers shooting at the optimal time point to capture the image of the object being photographed (S130).

[0049] In a preferred embodiment of the present invention, the motion sensing sensor 110 further includes a lidar and a gyroscope to realize a multi-sensor fusion scheme. In uniform motion scenarios with high speed, low light, or complex environments, traditional vision sensor-based systems suffer from large response delays and low positioning accuracy. This embodiment proposes a multi-sensor fusion scheme based on lidar and gyroscope to improve the robustness and accuracy of the system in complex environments. In this embodiment, the lidar sensor adopts the TOF (Time of Flight) principle, calculating the distance between the object and the camera 130 by emitting laser pulses and receiving reflected signals. This sensor has the characteristics of high precision, strong anti-interference ability, and fast response speed, and is suitable for complex environments such as low light and high-speed motion. The gyroscope sensor is used to measure the motion direction and angle changes of the object, which can effectively identify the attitude changes of the object, thereby correcting the imaging deviation caused by tilt or offset. The central processing unit 120 calculates a more accurate time-lapse shooting parameter B by fusing the data from the lidar and gyroscope sensors and combining them with a preset physical model. Camera 130 performs shooting operations under the control of central processing unit 120 and enables image caching mechanism to ensure clear images are captured in high-speed motion scenes. The technical advantages of this embodiment are: (1) multi-sensor fusion improves the system's adaptability in complex environments; (2) TOF ranging technology has higher accuracy and anti-interference capability than traditional visual ranging; (3) gyroscope sensor effectively corrects imaging deviations caused by changes in object posture; (4) image caching mechanism is suitable for high-speed uniform motion scenes, improving the success rate of image capture.

[0050] In practical applications, when the object being photographed enters the monitoring area, the LiDAR sensor first measures the distance between the object and the camera 130, while the gyroscope sensor simultaneously measures the object's direction of motion and angular changes. The central processing unit 120 fuses the data from the LiDAR sensor and the gyroscope sensor using a Kalman filter algorithm to obtain more accurate position and attitude information. Based on this information, the central processing unit 120 calculates the time delay shooting parameter B and triggers the camera 130 to capture images at the optimal moment. The camera 130 uses a high-speed continuous shooting mode, capturing 35 frames per second and buffering these images in its internal memory. When a trigger signal is received, the camera 130 selects the image closest to the trigger moment from the buffer to ensure that a clear image of the object is captured.

[0051] Figure 8 This is a flowchart illustrating an exemplary process of data acquisition and transmission by a motion sensing sensor 110 according to some examples of this application.

[0052] exist Figure 8 In the example, the position acquisition sensor 111 acquires the initial position of the object being photographed and generates a position-related analog signal (S111). Specifically, the position acquisition sensor 111 captures an image of the object being photographed through its internal visual sensor (such as a camera), and calculates the object's position based on image processing algorithms (such as edge detection and feature matching). Specifically, the visual sensor calculates the distance between the object and the sensor using formula (1) based on the principle of perspective projection, thereby determining the initial position. Here, the position acquisition sensor 111 is part of the motion sensing sensor 110.

[0053]

[0054] Where d is the distance between the object being photographed and the vision sensor, f is the focal length of the camera, L is the actual size of the object being photographed, and p is the pixel size of the object being photographed in the image. Furthermore, through the above calculations, the vision sensor can determine the initial position of the object being photographed.

[0055] exist Figure 8 In the example, the time acquisition sensor 112 records the initial time t0 when the object being photographed enters the detection area and generates a time-related analog signal (S112). Specifically, when the object being photographed enters the detection range of the motion sensing sensor 110, the time acquisition sensor 112 triggers an internal timer and records that moment as the initial time t0. The system 100 supports user-defined time settings, such as using the "00:00:00" format, or automatically synchronizing the time in different time zones through network connectivity to ensure global consistency of time information.

[0056] exist Figure 8In the example, the analog-to-digital converter 113 further converts the analog signals output by the position acquisition sensor 111 and the time acquisition sensor 112 into digital signals (S113). Specifically, the analog signals are sampled and quantized by the analog-to-digital converter (ADC) to convert continuous analog signals (such as voltage or current signals) into discrete digital signals (such as binary code). The converted digital signals are transmitted to the central processing unit 120 through the GPIO interface 114.

[0057] exist Figure 8 In the example, GPIO interface 114 transmits the digital signal after analog-to-digital conversion to the central processing unit 120 via a protocol (S114). GPIO interface 114 supports multiple communication protocols (such as I2C, SPI, and UART), and can select the appropriate protocol for data transmission according to different scenario requirements. In this embodiment, GPIO interface 114 adopts an interrupt response mechanism; that is, when the digital signal is ready, GPIO interface 114 sends an interrupt signal to the central processing unit 120, which immediately responds and reads the data, ensuring real-time data transmission. Furthermore, GPIO interface 114 also supports multi-channel parallel transmission, further improving data transmission efficiency and meeting the needs of high-speed scenarios.

[0058] Figure 9 This is a flowchart illustrating an exemplary process by which a central processing unit 120 runs an error compensation algorithm according to some examples of this application.

[0059] exist Figure 9 In the example, based on the initial position and velocity data collected by the motion sensing sensor 110, the central processing unit 120 calculates the basic delay parameters (S121) including spatial weight matrix compensation; the specific calculation formula (2) is as follows:

[0060] The basic delay parameter is calculated using spatial weighting:

[0061]

[0062] In formula (2), B base Defined as the fundamental theoretical delay parameter of the time-lapse photography control system. ⊙ represents the Hadamard product operation, realizing the sensor spatial weight matrix W. s The weight matrix, multiplied element-wise with the motion parameters, has diagonal elements ranging from 0.98 to 1.02, specifically designed to compensate for lens distortion and mechanical mounting errors. S represents the preset calibration distance from the motion sensing sensor 110 to the camera 130, measured in meters to ensure calculation accuracy. V is the instantaneous velocity vector of the object being photographed, obtained through continuous detection by the sensor 110.

[0063] exist Figure 9 In the example, the central processing unit 120 performs frequency domain dynamic compensation calculation on the basic delay parameters, and uses Fourier transform to decouple motion information from system delay, obtaining the frequency-domain corrected dynamic delay parameters (s122); frequency domain dynamic compensation achieves signal decoupling through Fourier transform:

[0064]

[0065] In formula (3), B freq Defined as the frequency-domain corrected dynamic delay parameter, this calculation process involves converting the velocity V and distance S to the frequency domain, where... This represents the Fourier transform operation. (*) denotes the inverse Fourier transform, and (*) denotes the frequency domain convolution operation. Δt sys This is the system's inherent delay parameter, ranging from 0.1 to 5 ms, and includes three main components: sensor response delay, signal processing delay, and mechanical execution delay. Through frequency domain feature interaction, this formula can effectively capture the dynamic correlation between motion speed and calibration distance.

[0066] Furthermore, the central processing unit 120 performs robust noise suppression correction on the dynamic delay parameters, using a corrected linear unit to filter out negative noise interference and introducing a Gaussian distribution to calibrate the noise term, outputting the final delay parameter B (S123); a noise suppression mechanism is introduced in the robust correction stage:

[0067]

[0068] In formula (4), the ReLU operator (·) corrects the linear unit, ensuring a positive output value and effectively filtering out negative noise interference. 2 The system signal-to-noise ratio variance reflects the level of environmental interference.

[0069] Furthermore, ε cal To calibrate the noise term (which follows a Gaussian distribution), the mathematical expansion is: ε cal ~N(0,δ 2 ), where δ 2 To calibrate the variance of the noise, this is used to simulate the random errors remaining during the hardware calibration process. In this embodiment, if the calibration process has been completed and the error characteristics are fixed (such as sensor factory calibration), then δ 2 Given the value ε cal It can be regarded as a random variable with a known distribution.

[0070] exist Figure 9In the example, camera 130 performs the first shooting action, and central processing unit 120 records the actual shooting time t1 and compares it with the theoretical value, calculates the time-varying noise function including sensor jitter and clock drift, and obtains the accurate time deviation Δt(s124); at this time, the deviation Δt between the actual shooting time and the theoretical value is calculated by formula (5):

[0071] Δt=t1-t0-B+ε(t) (5)

[0072] Where t0 records the initial time when the object enters the detection area, t1 is the actual shooting time of camera 130, and ε(t) is a time-varying noise function that includes sensor jitter and clock drift. Through this deviation detection mechanism, the system can continuously monitor and record the time error of each shot.

[0073] Specifically, the error compensation algorithm 1200, based on feedback from multiple shooting results, gradually corrects the deviation between the theoretical parameter B and the actual shooting time, ensuring that the position of the photographed object in the image is accurate and clear.

[0074] The error acquisition module 121 acquires the positional deviation Δx and time deviation Δt of the photographed object, which are used to provide initial data for the error compensation algorithm 1200 (S121). Specifically, based on multiple shooting results, the error acquisition module 121 analyzes the positional deviation Δx (i.e., the distance of the photographed object from the center position of the image acquired by the camera) and the time deviation Δt (i.e., the difference between the actual shooting time t1-t0 and the theoretical time delay shooting time parameter B). The positional deviation Δx and the time deviation Δt together constitute the basic data for error compensation, which are used for subsequent error modeling and parameter correction.

[0075] The error modeling module 122 establishes a correlation model between positional deviation Δx and time deviation Δt, providing a theoretical basis for the parameter correction module 123 (S122). Specifically, assuming the speed of the object being photographed is v, the relationship between positional deviation Δx and time deviation Δt can be expressed as the formula: Δx = v·Δt, that is, positional deviation Δx is equal to the product of the object's speed v and time deviation Δt, used to convert positional deviation into time deviation, further providing data support for the parameter correction module 123. Furthermore, through the error modeling module 122, the timed photography error compensation system further correlates positional deviation with time deviation, providing a basis for dynamically adjusting the time-lapse shooting parameter B. The error modeling module establishes a deviation model based on the formula Δx = v·Δt, and the parameter correction module uses the least squares method for iterative optimization.

[0076] The parameter correction module 123 dynamically adjusts the time-lapse shooting parameter B to generate a corrected parameter B′ (S123). Specifically, an iterative algorithm (such as the least squares method) is used to compensate for the acquisition accuracy of the initial time t0, and a corrected time-lapse shooting parameter B′ is generated. The calculation formula is: B′=B+k·Δt, that is, based on the original parameter B, the corrected parameter B′ is calculated by combining the time deviation Δt and the error compensation coefficient k. After each shooting, the system automatically updates the parameter B and applies the corrected parameter B′ to the subsequent shooting process. Through the parameter correction module 123, the system can gradually reduce the deviation between the theoretical parameter B and the actual shooting time t1-t0, so that the imaging position of the photographed object is gradually corrected to the center position of the camera lens 130.

[0077] Furthermore, the central processing unit 120 adopts a hybrid deep learning architecture including LSTM and Transformer, and dynamically adjusts the delay parameters by combining historical error data sequences, and outputs the optimized parameters B' (S126).

[0078] Specifically, the parameter correction module 123 uses a hybrid architecture deep learning approach to dynamically adjust the time-lapse shooting parameter B, which combines the multi-head attention mechanism of LSTM (Long Short-Term Memory Network) and Transformer, and is supplemented by gated linear units (GLU) and layer normalization (LayerNorm) techniques.

[0079] The corrected time-lapse shooting time parameter B' is calculated using formula (7):

[0080] B' = LayerNorm(B + GLU(W) c [Δt enh ;Δx])) (7)

[0081] In formula (7), LayerNorm represents the layer normalization operation, used to stabilize the training process and accelerate convergence; GLU(·) is a gated linear unit, which enables selective filtering of features; W c The weight matrix of the trainable GLU gated unit, with dimensions automatically adapted based on the input features; Δt enh To enhance time bias, deep features are extracted through nonlinear transformation; [Δt] enh ; Δx] indicates that the time deviation and spatial deviation are concatenated along the feature dimension.

[0082] Among them, the enhanced time deviation Δt enh Formula (8) yields the following:

[0083] Δt enh =PReLU(Δt) enh)+LSTM(Δt history (8)

[0084] In formula (8), PReLU is a parameterized modified linear unit, which has a stronger nonlinear expression capability compared to the standard ReLU function; LSTM is a long short-term memory network, specifically designed for processing time-series data; Δt history The mathematical expression for the historical time deviation sequence continuously recorded by the system is: Δt history ={Δt (1) , Δt (2) , …, Δt (k)}, where each element The object detection time Δt1 during the k-th capture was fully recorded. (k) Actual triggering time and the delay parameter B used (k) This sequence is maintained using a sliding window mechanism to ensure that the system always makes decisions based on the latest and most valid data.

[0085] Understandably, formula (7), as the core calculation module for the dynamic compensation parameter B' in the time-lapse photography control system, integrates three key technologies: residual learning, gating mechanism, and layer normalization. Its design achieves three important functions: first, by retaining the effective information of the original parameter B through residual connections, it avoids gradient vanishing; second, by utilizing the gating mechanism to achieve adaptive feature selection, it enhances the model's expressive power; and finally, by stabilizing the training process through layer normalization, it improves the model's convergence speed. This design enables the system to simultaneously handle time deviation Δt. enh It maintains excellent compensation performance in complex motion scenarios, including spatial deviation Δx.

[0086] The LSTM network in Equation (8) is specifically optimized for modeling time-series errors. The network employs a two-layer structure with 32 hidden layer units and uses the tanh activation function. During training, a teacher-forcing strategy is used, employing real historical errors as input to minimize the root mean square error (RMSE) between the predicted and actual errors. This network effectively captures the long-term dependencies of time biases, predicts potential future error trends, and provides forward-looking guidance for real-time compensation.

[0087] Figure 10 This is a flowchart illustrating an exemplary process by which a camera 130 triggers a shooting process according to some examples of this application.

[0088] exist Figure 10In the example, the camera 130 captures an image of the object being photographed upon triggering by the central processing unit 120 (S131). Specifically, the central processing unit 120 calculates the delayed shooting time parameter B based on the position and time information provided by the motion sensing sensor 110, and triggers the camera 130 to capture the image at the initial optimal time point T. More specifically, the central processing unit 120 calculates the initial optimal time point T using the formula: T = t0 + B, that is, the initial optimal time point T is equal to the sum of the initial time t0 and the delayed shooting time parameter B. After receiving the trigger signal, the camera 130 immediately performs the shooting operation at time point T, captures an image of the object being photographed, and transmits the image data to the central processing unit 120 for subsequent processing.

[0089] exist Figure 10 In the example, the triggering mechanism of camera 130 works in conjunction with the error compensation algorithm 1200 of central processing unit 120 to dynamically adjust the shooting time point to improve image capture accuracy, ultimately ensuring that the image is clear and the photographed object is located in the center of the image (S132). Specifically, during multiple shooting processes, the central processing unit 120 dynamically adjusts the delayed shooting time parameter B through the error compensation algorithm 1200, generates a corrected parameter B′, and applies the corrected parameter B′ to the triggering logic of camera 130. More specifically, the central processing unit 120 calculates the corrected optimal time point T′, and the formula for calculating the corrected optimal time point T′ is: T′=t0+B′, that is, the corrected optimal time point T′ is equal to the sum of the initial time t0 and the corrected delayed shooting time parameter B′. When a positional deviation Δx of the photographed object in the image is detected, the central processing unit 120 adjusts the triggering time according to the deviation data, so that camera 130 can shoot at the corrected optimal time point T′, further gradually correcting the imaging position of the photographed object to the center position of the camera 130 lens.

[0090] This application provides a method for timed image capture error compensation in uniform motion scenarios. A motion sensing sensor 110 collects the initial position and initial time of the object being photographed in real time. A central processing unit 120 dynamically adjusts the delayed shooting time parameters based on an error compensation algorithm 1200 and precisely controls the triggering timing of the camera 130, ensuring the accurate and clear position of the object in the image. This system, through the collaborative work of multiple modules, achieves end-to-end optimization from data acquisition and error compensation to image capture, improving shooting hit rate and image accuracy. It is applicable to various scenarios such as industrial production, logistics sorting, and traffic monitoring, meeting the needs of high-precision image capture.

[0091] The above embodiments can be implemented entirely or partially by software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented entirely or partially as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0092] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0093] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0094] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0095] The logic and / or steps described in the flowchart or otherwise herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0096] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, the program being stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiments.

[0097] The foregoing description and illustrations have shown and described the basic principles, main features, and advantages of this invention. Those skilled in the art should understand that this invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to this invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A timed photography error compensation system for uniform motion scenes, characterized in that, include: Motion sensing sensors, a central processing unit, and at least one camera. The motion sensing sensor is used to collect the initial position and initial time of the object being photographed and generate corresponding signals. The motion sensing sensor converts the collected analog signals into digital signals and transmits them to the central processing unit through the GPIO interface of its internal processor. The central processing unit is used to receive the digital signal and calculate the time-lapse shooting parameters, wherein the time-lapse shooting parameters are calculated based on the distance and relative speed between the photographed object and the camera; the central processing unit is also used to run an error compensation algorithm to dynamically adjust the time-lapse shooting parameters based on multiple shooting results in order to optimize the position of the photographed object in the image; A camera is used to capture an image of the object being photographed when triggered by the central processing unit; wherein the error compensation algorithm includes an association model based on time deviation and position deviation, and uses a sliding window weighted average method to dynamically correct the time delay shooting parameters.

2. The timed photography error compensation system according to claim 1, characterized in that: The motion sensing sensor includes a position acquisition sensor, a time acquisition sensor, and an analog-to-digital converter. The position acquisition sensor and the time acquisition sensor are used to acquire the initial position and initial time of the object being photographed, respectively, and the acquired analog signals are converted into digital signals by the analog-to-digital converter.

3. The timed photography error compensation system according to claim 2, characterized in that: The GPIO interface supports multiple communication protocols and triggers the central processing unit to read data through an interrupt response mechanism, ensuring the real-time performance of data acquisition and processing.

4. The timed photography error compensation system according to claim 1, characterized in that: The central processing unit includes: The error acquisition module is used to collect the positional and temporal deviations of the photographed object; The error modeling module is used to establish a correlation model between positional deviation and time deviation; The parameter correction module is used to dynamically adjust the time-lapse shooting parameters; wherein, the error modeling module establishes a deviation model based on the formula Δx=v·Δt, where Δx is the positional deviation, v is the uniform velocity of the object being photographed, and Δt is the time deviation.

5. The timed photography error compensation system according to claim 4, characterized in that: The parameter correction module uses the least squares method for iterative optimization to generate corrected time-lapse shooting parameters, gradually converging to the optimal shooting time point.

6. The timed photography error compensation system according to claim 4, characterized in that: The camera supports multiple shooting modes, including single shooting, continuous shooting, and timed shooting, and performs shooting operations under the control of the central processor to ensure that objects in the image are in the optimal position.

7. A method for compensating for timing-based photography errors in uniform motion scenarios, characterized in that, Includes the following steps: The initial position and initial time of the object being photographed are collected by the motion sensing sensor, a digital signal is generated and transmitted to the central processing unit via the GPIO interface; The central processing unit receives the digital signal and calculates the time-lapse shooting parameters based on the distance and relative speed between the object being photographed and the camera. The central processing unit runs an error compensation algorithm to adjust the time delay shooting parameters based on multiple shooting results, thereby optimizing the position of the photographed object in the image. The camera captures an image of the object under the trigger of the central processing unit; wherein the error compensation algorithm includes an association model based on time deviation and position deviation, and uses a sliding window weighted average method to dynamically correct the time delay shooting parameters.

8. The timed photography error compensation method according to claim 7, characterized in that: The error modeling module establishes a deviation model based on the formula Δx=v·Δt, where Δx is the position deviation, v is the uniform velocity of the photographed object, and Δt is the time deviation. The parameter correction module uses the least squares method for iterative optimization to generate corrected time-lapse shooting parameters.

9. The timed photography error compensation method according to claim 7, characterized in that: The central processing unit determines the initial optimal shooting time point based on the time-lapse shooting time parameter, and triggers the camera to perform a shooting operation at the initial optimal shooting time point; After multiple shots, the time-lapse shooting parameters are dynamically adjusted based on the actual shooting results to determine the optimal shooting time point, and the camera is triggered to perform a shooting operation at the optimal shooting time point.

10. The timed photography error compensation method according to claim 7, characterized in that: The error compensation algorithm further includes adopting a hybrid deep learning architecture of LSTM and Transformer, combining historical error data sequences to dynamically adjust the delay parameters, and outputting optimized parameters B' to improve the compensation performance of the system in complex motion scenarios.

Images