A non-contact item size measurement system based on edge computing

By using a non-contact object size measurement system based on edge computing, distortion and color correction are performed using ordinary cameras and edge computing units, combined with a target detection model. This solves the problems of low efficiency, high cost, and poor real-time performance in existing technologies, and achieves efficient, low-cost, and real-time object size measurement.

CN224984801UActive Publication Date: 2026-11-17HUMMINGBIRD INNOVATION (BEIJING) TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202522810130.8
Authority / Receiving Office
CN · China
Patent Type
Utility models(China)
Current Assignee / Owner
Filing Date
2025-12-30
Publication Date
2026-11-17
Estimated Expiration
2035-12-30

AI Technical Summary

Technical Problem

Existing technologies for measuring the dimensions of items are inefficient, costly, lack real-time performance, and have poor environmental adaptability, making it difficult to meet the automation and real-time requirements of scenarios such as manufacturing and logistics warehousing.

Method used

A non-contact object size measurement system based on edge computing is adopted. It uses a regular monocular camera combined with an edge computing unit for image processing to achieve distortion and color correction. Combined with a lightweight target detection and key point detection model, it can perform efficient and accurate size measurement.

Benefits of technology

It achieves efficient, low-cost, and real-time measurement of item dimensions, increasing measurement speed by 150 times, achieving an accuracy of less than 1%, reducing human error and hardware costs, and adapting to various environmental changes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN224984801U_ABST
    Figure CN224984801U_ABST
Patent Text Reader

Abstract

This application discloses a non-contact object size measurement system based on edge computing, including a working platform, a vertical support mounted on the working platform, and a camera mounting plate and an integrated mounting box mounted on the vertical support. The upper surface of the working platform is marked with a test object identification position and multiple size identification positions. The test object identification positions are used to place color cards or test objects, and the size identification positions are used to place markers of known sizes. The integrated mounting box is equipped with an edge computing unit for obtaining calibration parameters or object sizes based on the acquired corresponding images. This system can achieve non-contact measurement of the size of test objects, with fast measurement speed and high accuracy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the fields of computer vision and industrial automation technology, and in particular to a non-contact object size measurement system based on edge computing. Background Technology

[0002] In manufacturing, process inspection, logistics warehousing, and retail settlement, it is frequently necessary to measure the dimensions of items. Existing item dimension measurement technologies have several shortcomings, primarily including:

[0003] 1. Low efficiency and automation: Traditional contact measurement (such as tape measure and caliper) relies on manual labor, is slow, prone to errors, and difficult to integrate into automated production lines.

[0004] 2. High cost and deployment threshold: High-precision non-contact solutions (such as 3D LiDAR and binocular depth cameras) have high hardware costs, complex equipment, and are difficult to deploy and maintain.

[0005] 3. Insufficient real-time performance and reliability: Cloud-based solutions suffer from high network latency and bandwidth consumption, are susceptible to network fluctuations, make it difficult to meet real-time requirements, and pose data security risks.

[0006] 4. Poor environmental robustness: Ordinary vision solutions are easily affected by lens distortion and changes in lighting, resulting in large measurement errors and lacking the ability to adaptively calibrate during long-term operation. Utility Model Content

[0007] In order to address the problems of low efficiency, high cost, poor real-time performance, and weak environmental adaptability in existing item size measurement technologies, this application is proposed to provide a non-contact item size measurement system based on edge computing that overcomes or at least partially solves the above problems.

[0008] This application provides a non-contact item size measurement system based on edge computing, including: a work platform, a vertical bracket disposed on the work platform, a camera mounting plate for suspending a camera mounted on the top of the vertical bracket, and an integrated mounting box mounted on the vertical bracket.

[0009] The upper surface of the work platform is marked with a test object identification position and multiple size identification positions. The test object identification position is used to place a color card or the test object, and the size identification positions are used to place a calibration object of known size.

[0010] The integrated mounting box is equipped with an edge computing unit, which is used to obtain calibration parameters or item dimensions based on the acquired corresponding images.

[0011] Optionally, the center of the test object identifier coincides with the center of the upper surface of the working platform.

[0012] Optionally, the object to be tested identifier is rectangular, with its center located directly below the camera mounting plate.

[0013] Optionally, the upper surface of the work platform is rectangular, and each of its four corner areas is provided with a size mark, and each size mark is a circle of the same size.

[0014] Optionally, the vertical support is inverted L-shaped, with the top of the inverted L-shaped vertical support extending toward the working platform.

[0015] Optionally, the integrated mounting box is installed on the side of the vertical bracket opposite to the working platform.

[0016] Optionally, the integrated mounting box is also equipped with a main control unit for controlling the camera's shooting parameters.

[0017] Optionally, a checkerboard calibration plate is also included, which is placed on the upper surface of the work platform when the system is in the distortion calibration working state.

[0018] Optionally, when the system is in color calibration mode, the test object identifier is used to place the color card;

[0019] When the system is in the dimension calibration working state, the dimension marker is used to place a calibration object of known size;

[0020] When the system is in the size measurement working state, the object identification bit is used to place the object to be measured.

[0021] Optionally, the bottom of the work platform is provided with four support pads or four sliding wheels with braking mechanisms.

[0022] The beneficial effects of the above-mentioned technical solutions provided in the embodiments of this application include at least the following:

[0023] (1) The edge computing-based non-contact item size measurement system provided in this application embodiment performs non-contact size measurement based on images captured by a camera. Compared with manual measurement, it reduces human error factors, avoids operational damage to the item to be measured that may be caused by manual measurement, and improves measurement speed. After actual testing, based on the fact that there are 10 measurement positions for each item, the average time for manual measurement and recording of smaller items is about 1 minute, and the average time for larger items is about 5 minutes. However, the edge computing-based non-contact item size measurement system provided in this application embodiment can shorten the overall measurement and recording time to 2 seconds, which can improve efficiency by up to 150 times.

[0024] (2) The edge computing-based non-contact object size measurement system provided in this application embodiment utilizes distortion correction and color correction to obtain high-precision and reliable size results on low-cost hardware devices. The system can use a common monocular camera, without the need for a high-end camera with extremely low distortion and extremely high resolution.

[0025] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0026] The technical solution of this application will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0027] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the embodiments of the present application to explain the application and do not constitute a limitation thereof. In the drawings:

[0028] Figure 1 This is a three-dimensional structural diagram of the edge computing-based non-contact article size measurement system provided in the embodiments of this application;

[0029] Figure 2 This is a front view structural diagram of the non-contact article size measurement system based on edge computing provided in the embodiments of this application;

[0030] Figure 3 This is a top view of the non-contact article size measurement system based on edge computing provided in this embodiment of the application.

[0031] Figure 4 This is a system software logic design diagram provided in the embodiments of this application;

[0032] Figure 5 This is a calibration flowchart provided in the embodiments of this application;

[0033] Figure 6 This is a measurement flowchart provided in the embodiments of this application. Detailed Implementation

[0034] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0035] In the description of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," "outer," "far," "near," "front," and "rear," etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the accompanying drawings and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this application. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0036] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0037] Example 1

[0038] This application discloses a non-contact item size measurement system based on edge computing. See [link to relevant documentation] Figures 1-3 As shown, it includes:

[0039] (1) Working platform 1: Provides the measurement plane of the item to be measured and the calibration plane before measurement.

[0040] (2) Vertical support 2: set on the work platform 1, and fixedly or detachably connected to the work platform 1.

[0041] (3) Integrated mounting box 3: installed on the vertical bracket 2, specifically, it can be installed on the side of the vertical bracket 2 away from the working platform 1.

[0042] The integrated mounting box 3 is equipped with an edge computing unit for obtaining calibration parameters or item dimensions based on the acquired corresponding image.

[0043] Calibration parameters include camera intrinsic matrix A, distortion coefficient vector D, color correction matrix CCM, and pixel equivalent (mm / pixel).

[0044] The specific calibration parameters and the calculation of the item dimensions can be achieved using existing methods.

[0045] (4) Camera mounting plate 4: Installed on the top of the vertical bracket 2 for hanging the camera. The camera mounting plate 4 is designed to be compatible with different models of cameras.

[0046] A standard monocular camera can be used, such as a regular industrial USB camera with a resolution of 1920×1080.

[0047] The upper surface of the work platform 1 is marked with a test object identification position 5 and multiple size identification positions 6. The test object identification position 5 is used to place a color card (such as a Macbeth 24 color card) or the test object, and the size identification positions 6 are used to place a calibration object with a known size.

[0048] The system also includes a checkerboard calibration plate, which is placed on the upper surface of the work platform when the system is in distortion calibration mode. Specifically, it can be placed sequentially at the lower left, upper left, upper center, lower center, lower right, and upper right of the upper surface of the work platform to ensure optimal calibration results.

[0049] When the system is in color calibration mode, the test object identifier is used to place the color card; when the system is in size calibration mode, the size identifier is used to place the calibration object of known size; when the system is in size measurement mode, the test object identifier is used to place the test object.

[0050] In some embodiments, the integrated installation box may be an integrated industrial control computer (x86 architecture) as the main control unit, and an integrated NVIDIA Jetson NX as the edge computing unit, responsible for running image processing and deep learning algorithms.

[0051] Optionally, the software configuration of the edge computing unit is as follows:

[0052] S1: Deploy the Ubuntu operating system on the edge computing unit and install software environments such as Python, OpenCV, and PyTorch.

[0053] S2: Use Zhang Zhengyou's calibration method to write a calibration script, control the camera to capture multiple chessboard images, calculate and save the camera intrinsic parameter matrix A and distortion coefficient vector D.

[0054] S3: Write a color correction script to extract the RGB values ​​of the 24 color card area in the calibration board, fit them with the standard values ​​to obtain a 3x3 color correction matrix CCM, and save it.

[0055] S4: Write a scale calculation script to identify four circular markers in the image, calculate the average pixel equivalent (mm / pixel) based on their known physical diameter (e.g., 50mm) and pixel diameter, and save the result.

[0056] S5: Deploy the lightweight object detection model YOLOV8s and the key point detection model VITPose to the edge computing unit.

[0057] The above software configurations are all based on existing technologies. Optional, other existing software configurations can also be used, as long as they can achieve the relevant functions of parameter calibration and item size measurement.

[0058] In some embodiments, the center of the test object identifier 5 coincides with the center of the upper surface of the working platform 1.

[0059] Furthermore, the object to be tested identifier 5 is rectangular, with its center located directly below the camera mounting plate 4.

[0060] In some embodiments, the upper surface of the work platform 1 is rectangular, and a size identifier 6 is provided at each of its four corner regions. The distance between each size identifier 6 and the long side of the nearest rectangle (the upper surface of the work platform), and the distance between each size identifier 6 and the short side of the nearest rectangle, can both be the same.

[0061] Furthermore, each size identifier 6 is a circle of the same size.

[0062] In some embodiments, the vertical support 2 is inverted L-shaped, and the top of the inverted L-shaped vertical support 2 extends toward the working platform 1.

[0063] The vertical bracket 2 is telescopic, so the distance between the camera suspended at its top and the upper surface of the work platform 1 is adjustable.

[0064] The bottom of the work platform 1 is equipped with four support pads to prevent friction. Alternatively, four sliding wheels with braking mechanisms can be installed at the bottom of the work platform.

[0065] The integrated installation box 3 also includes a main control unit responsible for business scheduling. This includes a camera control unit for controlling camera shooting parameters; a calibration process unit for scheduling distortion / color correction and exposure adaptation; and an interaction unit for human-computer interaction and result display.

[0066] The above settings for the main control unit can be achieved using the conventional main control settings for cameras in existing technologies.

[0067] The edge computing-based non-contact object size measurement system provided in this application embodiment performs non-contact size measurement based on images captured by a camera. Compared with manual measurement, it reduces human error factors, avoids potential damage to the object being measured due to manual measurement, and improves measurement speed. Actual testing, with 10 measurement locations per object, showed that manual measurement and recording time for smaller objects averaged around 1 minute, while for larger objects it averaged around 5 minutes. The edge computing-based non-contact object size measurement system provided in this application embodiment reduces the overall measurement and recording time to 2 seconds, achieving an efficiency improvement of up to 150 times.

[0068] The edge computing-based non-contact object size measurement system provided in this application embodiment utilizes distortion correction and color correction to achieve high-precision and reliable size results on low-cost hardware. This system can use a standard monocular camera, eliminating the need for high-end cameras with extremely low distortion and extremely high resolution.

[0069] Example 2

[0070] Embodiment 2 of this application provides a non-contact article size measurement method based on edge computing, which uses the measurement system provided in Embodiment 1 to measure article size, and includes the following steps:

[0071] (1) System installation

[0072] Level and calibrate the working platform 1, and install the vertical bracket 2 on the working platform 1; then install the camera mounting plate 4 and the integrated mounting box 3 onto the vertical bracket 2 respectively; then install the camera 7 onto the camera mounting plate 4. After the installation and wiring are completed, begin the calibration work before measurement.

[0073] (2) System initialization and calibration

[0074] When using the system for the first time or when environmental conditions change drastically, this procedure needs to be performed for calibration to reduce measurement errors caused by changes in external parameters such as camera and environment.

[0075] a. Distortion correction: The checkerboard calibration plate is placed sequentially on the lower left, upper left, upper center, lower center, lower right, and upper right of the working platform surface. By taking checkerboard images at multiple positions, the camera intrinsic parameters and distortion coefficients are calculated and stored in the edge computing unit.

[0076] b. Color Correction: By capturing images of a standard color chart, the color correction matrix (CCM) is calculated and stored in the edge computing unit.

[0077] c. Size ratio calculation: By shooting a scene with known size reference markers, calculate the conversion ratio of pixels to physical units (such as millimeters) in the current field of view, and store it in the edge computing unit.

[0078] The above calibration process can be executed during system deployment or triggered periodically.

[0079] (3) Adaptive image acquisition and preprocessing:

[0080] a. The camera captures the original image containing item 8 to be tested.

[0081] b. The edge computing unit performs rapid analysis on the original image (such as assessing the overall brightness of the object area), dynamically adjusts the camera exposure parameters, and obtains optimized image quality with appropriate exposure.

[0082] c. Using the stored calibration parameters, perform distortion correction and color correction on the optimized image in sequence, analyze the average brightness of the corrected image, and if it deviates from the preset range, automatically adjust the camera exposure parameters and take a new picture. Then repeat the distortion and color correction on the new picture to obtain the corrected image to be analyzed.

[0083] (4) Item recognition and key point extraction:

[0084] a. Input the corrected image into the object detection model, such as the YOLOV8s model, to obtain the precise bounding box of the object.

[0085] b. Based on this bounding box, input a key point detection model (such as the VITPose model). The model first predicts the approximate location of the key points of the object, and then performs fine regression near each key point to output high-precision key point pixel coordinates.

[0086] (5) Dimension calculation and result output:

[0087] a. Calculate the pixel distance of the dimension to be measured on the image based on the pixel coordinates of the key points.

[0088] b. Call the stored pixel-to-physical size scale to convert the pixel distance into the actual physical size (length, width, etc.).

[0089] c. Draw key points and dimension annotations on the image, and then transmit the resulting data (dimensional values, image) to the main control module.

[0090] Using the above embodiments, standardized items of various specifications were measured. The dimensional measurement results were compared with those measured by manual calipers. The average error range was 1mm-5mm. The total time for a single measurement (from taking a picture to displaying the result) was stable within 2000ms. This achieved automated measurement with high precision (error ≤1%), low latency (≤2000ms), and low cost.

[0091] The specific implementation methods of the above steps (2)-(5) can all be achieved by using mature methods in the existing technology.

[0092] See Figure 4The diagram shows the overall system design, which includes three main modules: a camera module, a main control unit, and an edge computing unit. The main control unit comprises an image acquisition module and a parameter calibration module, while the edge computing unit includes distortion correction algorithms, color correction algorithms, HSV recognition algorithms, edge detection algorithms, object detection algorithms, and keypoint algorithms. The main control unit is responsible for overall coordination and control throughout the system, calling upon the relevant functions of the camera module and the edge computing unit to complete the overall business process.

[0093] See Figure 5 The diagram shows the calibration flowchart. This process is required when using the system for the first time or when environmental conditions change drastically to reduce dimensional measurement errors. The calibration process consists of three main parts: (a) distortion correction; (b) color correction; and (c) positioning point recognition. These three processes work together to reduce measurement errors caused by changes in external parameters such as camera and environment.

[0094] See Figure 6 The diagram shows the measurement flowchart. When preparing to measure an object, the system first automatically identifies key points and determines whether recalibration is needed. Then, it executes the normal detection process, including the HSV algorithm process, distortion correction process, color correction process, object detection process, key point recognition process, and size calculation process. The HSV algorithm is primarily responsible for identifying the object's color to adjust the camera exposure appropriately, preventing overexposure and underexposure, making the details and textures of the captured object image clearer, and improving the accuracy of the deep learning algorithm. Distortion correction and color correction mainly use the correction coefficients obtained during calibration to correct the image, reducing accuracy errors caused by hardware. Finally, object detection and key point detection identify the corresponding size key point information and perform size calculations to obtain relatively accurate size data.

[0095] Figure 5 and Figure 6 The calibration and measurement procedures can be implemented based on existing technologies.

[0096] Obviously, those skilled in the art can make various modifications and variations to this application without departing from its spirit and scope. This disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A non-contact object size measurement system based on edge computing, characterized in that, Includes: a work platform, a vertical bracket set on the work platform, a camera mounting plate for suspending the camera mounted on the top of the vertical bracket, and an integrated mounting box mounted on the vertical bracket; The upper surface of the work platform is marked with a test object identification position and multiple size identification positions. The test object identification position is used to place a color card or the test object, and the size identification positions are used to place a calibration object of known size. The integrated mounting box is equipped with an edge computing unit, which is used to obtain calibration parameters or item dimensions based on the acquired corresponding images.

2. The system as described in claim 1, characterized in that, The center of the test object identifier coincides with the center of the upper surface of the working platform.

3. The system as described in claim 1, characterized in that, The object to be tested is a rectangular marker, with its center located directly below the camera mounting plate.

4. The system as described in claim 1, characterized in that, The upper surface of the work platform is rectangular, and each of its four corner areas is provided with a size mark, and each size mark is a circle of the same size.

5. The system as described in claim 1, characterized in that, The vertical support is inverted L-shaped, with the top of the inverted L-shape extending toward the working platform.

6. The system as described in claim 1, characterized in that, The integrated mounting box is installed on the side of the vertical support opposite to the working platform.

7. The system as described in claim 1, characterized in that, The integrated mounting box is also equipped with a main control unit for controlling the camera's shooting parameters.

8. The system as described in claim 1, characterized in that, It also includes a checkerboard calibration plate, which is placed on the upper surface of the work platform when the system is in the distortion calibration working state.

9. The system as described in claim 8, characterized in that, When the system is in color calibration mode, the test object identifier is used to place the color card; When the system is in the dimension calibration working state, the dimension marker is used to place a calibration object of known size; When the system is in the size measurement working state, the object identification bit is used to place the object to be measured.

10. The system according to any one of claims 1 to 9, characterized in that, The bottom of the work platform is provided with four support pads or four sliding wheels with braking mechanisms.