A method and system for acoustic intelligent detection of empty containers

Abstract

The present invention provides an intelligent acoustic detection method and system for empty containers. The method comprises the following steps: S1: obtaining the size and position information of the container using a visual detection method, and accurately placing acoustic excitation and acoustic receiving devices at corresponding designated positions using an automatic placement device; S2: acoustically exciting the container using the acoustic excitation device, while simultaneously collecting the container's acoustic response signal using a microphone in the acoustic receiving device; S3: performing noise reduction processing on the response signal using a transfer function method to obtain a true spectrum estimate for each measurement point; S4: calculating the container's acoustic clustering signal using an acoustic clustering algorithm and extracting the container's acoustic cluster features; S5: comparing the extracted acoustic cluster features with the reference acoustic cluster features of the container identified in step S1 to achieve intelligent detection of the container's status. The present invention achieves intelligent empty container detection and also has the advantages of sensitive detection features, excellent stability, and anti-interference capabilities, thereby improving the accuracy of detection results.

Description

Technical Field

[0001] The present invention relates to the technical field of container detection, and in particular to an acoustic intelligent detection method for empty containers. Background Art

[0002] With the rapid development of my country's foreign trade, the task of container supervision during entry and exit has become increasingly arduous. Under the new technological revolution of industrial automation and intelligent technology, the demand for high-performance intelligent container supervision technology has become increasingly strong.

[0003] Current empty container detection technologies are categorized into three main categories based on their level of intelligence: The first is traditional, fully manual inspection and decision-making; the second is semi-intelligent methods that rely on automated cameras or X-rays for visual inspection and manual decision-making. For example, Patent (Granted Announcement No. CN 216927109U) uses a camera located on an elevator to photograph containers to determine if they are empty. Patent (Granted Announcement No. CN216013666U) and Patent (Granted Announcement No. CN217112754U) respectively provide methods for detecting empty containers at the head and / or tail ends using lidar and 3D laser scanners; Patent (Granted Announcement No. CN113409239B) proposes a method for empty container detection based on radiometric imaging. For the above two types of technologies, the detection efficiency is low, and there are serious problems such as low accuracy caused by human instability. Among them, the ray-type detection method also has problems such as high equipment cost and radiation hazards to personnel and the environment caused by the physical medium used for detection. The third type is a fully intelligent method that uses physical means such as acoustics and machine learning decision-making. For example, the patent (authorization announcement number CN 107621653B) uses knocking and / or transducers as the box excitation source, collects signals through acceleration sensors, and analyzes the reverberation, resonance frequency, coherence or any combination of the above inside the box to determine whether the container is empty or not. The patent (authorization announcement number CN 207751939U) provides a container detector that uses a robotic arm to move the empty box detection device to the bottom beam position of the box for excitation, and the collected box sound wave frequency can accurately determine whether the container is empty or has a mezzanine. The patent (authorization announcement number CN110231402B) provides a container detection method and device, which uses an air-mediated excitation signal to stimulate designated locations on the container, collects the composite vibration response signal on the container, and extracts the features to send to deep learning models such as neural networks for intelligent identification of the container's status. For the above-mentioned acoustic intelligent detection technology and products, the detection features used (such as resonance frequency, response attenuation rate, etc.) have the following defects in actual application: 1. The detection features have little difference between the empty and non-empty states, and the sensitivity is poor and unstable; 2. They are easily affected by uncontrollable factors such as environmental noise, local modes of the box wall panels, and differences in the box material. The above defects result in the actual detection accuracy of existing acoustic intelligent detection technologies being unguaranteed, and the use effect is poor. Summary of the Invention

[0004] In view of the deficiencies in the above-mentioned prior art, the purpose of the present invention is to provide an acoustic intelligent detection method and system for empty containers that has a high degree of intelligence and detection accuracy and can better meet actual usage needs.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] An acoustic intelligent detection method for empty containers comprises the following steps:

[0007] S1: Use visual detection methods to obtain the size and position information of the entry box, and use the automatic placement device to accurately place the acoustic excitation and acoustic receiving devices at the corresponding designated positions;

[0008] S2: The cabinet is acoustically excited by the acoustic excitation device, and the acoustic response signal of the cabinet is synchronously collected by the microphone in the acoustic receiving device;

[0009] S3: Use the transfer function method to perform noise reduction on the response signal to obtain the true spectrum estimation of each measurement point;

[0010] S4: Calculate the cabinet acoustic clustering signal through the acoustic clustering algorithm and extract the cabinet acoustic cluster features;

[0011] S5: Compare the extracted acoustic cluster features with the reference acoustic cluster features of the cabinet identified in step S1 to achieve intelligent detection and judgment of the cabinet status.

[0012] Preferably, the number of the acoustic excitation device in step S1 is 1, and its designated working position is located at a corner point of the box.

[0013] Preferably, in step S1, the number of microphones in the sound receiving device is multiple, one of which is a designated working position at the same corner point of the box as the position of the sound excitation device, and the remaining designated working positions are located at corner points of at least two box surfaces.

[0014] Preferably, the true spectrum estimation of each measuring point in step S3 is The calculation method is as follows

[0015]

[0016] in, For p i The cross spectrum between (f) and r(f), r(f) is the reference acoustic signal near the acoustic excitation device, p i (f) is the acoustic signal of the measuring points at other corner positions, C rr (f) = E[r * (f)r(f)], is the autospectrum of r(f), E[·] represents the mathematical expectation, i = 1, 2, …, K-1, K is the total number of microphones.

[0017] Preferably, the cabinet acoustic clustering algorithm in step S4 is calculated as follows:

[0018] The signal vectors of the four corner points on the surface of each box where the measuring points are arranged Use Hadamard matrix T to perform matrix transformation to obtain the cabinet acoustic clustering signal

[0019] Where T = (t1 t2 t3 t4), t1 = (1, 1, 1, 1), t2 = (1, 1, -1, -1), t3 = (1, -1, -1, 1), t4 = (1, -1, 1, -1).

[0020] Preferably, the acoustic cluster features of the box in step S4 are extracted by the following method:

[0021] Acoustic clustering signal of the jth box surface where measurement points are arranged Assume that the spectrum values ​​at the analysis frequencies k-1, k and k+1 are and Set the threshold ratio δ, record like and Zeji The cluster feature is 1, otherwise it is 0, thus obtaining the acoustic cluster feature S of the j-th box surface frequency point k j (k), and finally, the acoustic cluster features of the box surface where all measurement points are arranged are combined into the acoustic cluster features of the box [s1(k),…,s j (k)].

[0022] Preferably, in the above-mentioned intelligent acoustic detection method for empty containers, the intelligent detection and judgment of the container status in step S5 is performed as follows:

[0023] The acoustic cluster characteristics of the inspected box [S1(k),…, S j(k)] is recorded as vector a, and step S1 identifies the reference acoustic cluster features of the box Denote it as vector b, and calculate the cosine of the angle between them cos(a,b) =<a,b> / (||a||||b||), if cos(a,b)=0, it is judged to be an empty box state, otherwise it is a non-empty box state.

[0024] According to another aspect of the present invention, there is also provided an acoustic intelligent detection system for empty containers, which includes a visual detection module, an acoustic excitation and receiving module, a support platform including an automatic deployment device, a data acquisition and preprocessing module, and a computer control system including relevant algorithms;

[0025] The visual detection module is used to detect the size and position information of the box;

[0026] The acoustic excitation and receiving module is used to acoustically excite the box at a specified position and receive an acoustic response signal of the box at the specified position;

[0027] The support platform including the automatic deployment device is used to install equipment such as the visual detection module, the acoustic excitation and receiving module, and automatically deploy and recover the acoustic excitation and receiving module as required;

[0028] The data acquisition and preprocessing module is used to collect signals such as the cabinet acoustic response and perform preprocessing functions such as filtering;

[0029] The computer control system is used to execute all programs and algorithms such as visual inspection, automatic placement and recovery control, data collection and analysis processing, data reading and writing, and result output.

[0030] Due to the adoption of the above-mentioned scheme, the intelligent acoustic detection method and system for empty containers provided by the present invention utilize a transfer function method to perform noise reduction processing on the response signal of the container, and extract the acoustic cluster characteristics of the inspected container through a container acoustic clustering algorithm. The system has the advantages of sensitive detection characteristics, excellent stability and anti-interference ability, and improves the accuracy and stability of the detection results. The system utilizes a visual detection method to obtain the size and position information of the container, automatically deploys and recovers the acoustic excitation and sound receiving devices, and realizes automatic detection and judgment of the container status by extracting the acoustic cluster characteristics of the container and comparing them with the reference acoustic cluster characteristics. The system has a high degree of intelligence and is conducive to the promotion and application of actual detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic flow chart of an acoustic intelligent detection method for empty containers according to an embodiment of the present invention.

[0032] Figure 2 Schematic diagram of designated working positions of an acoustic excitation device and a microphone according to an embodiment of the present invention.

[0033] Figure 3 It is the calculation result of the acoustic clustering signal of the two measurement surfaces of the cabinet in the embodiment of the present invention.

[0034] Figure 4 This is a schematic diagram of the structure of the acoustic intelligent detection system for empty containers according to an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0036] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like, indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0037] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood broadly. For example, they may refer to fixed connections, removable connections, or integral connections. They may refer to mechanical connections or electrical connections. They may refer to direct connections or indirect connections through an intermediary, and they may refer to internal communication between two components or interactions between two components. Those skilled in the art will understand the specific meanings of these terms in the present invention based on the specific circumstances.

[0038] like Figures 1 to 4 As shown, Figure 1 The following is a flow chart of an acoustic intelligent detection method for empty containers provided by an embodiment of the present invention, comprising the following steps:

[0039] S1: Use visual detection methods to obtain the size and position information of the entry box, and use the automatic placement device to accurately place the sound excitation and sound receiving devices at the corresponding designated positions.

[0040] In this embodiment, if Figure 2 As shown in the figure, the number of acoustic excitation devices is 1, and its designated working position is located at a corner point of the box; the number of microphones in the sound receiving device is 7, one of which is a reference microphone, and its designated working position is the same corner point of the box as the position of the acoustic excitation device, and the remaining 6 are designated working positions located at various corner points on the surfaces of the two boxes.

[0041] S2: The cabinet is acoustically excited by the acoustic excitation device, and the acoustic response signal of the cabinet is synchronously collected by the microphone in the acoustic receiving device.

[0042] The acoustic excitation device used in this implementation is a high-intensity electroacoustic transducer that generates a pulsed acoustic signal with a sound pressure level exceeding 120dB at a designated operating position to excite the cabinet. Microphones located at two corners of the cabinet surface are pre-triggered by the signal amplitude of a reference microphone to synchronously collect the cabinet's acoustic response signals.

[0043] S3: Use the transfer function method to reduce the noise of the response signal and obtain the true spectrum estimation of each measurement point.

[0044] The presence of various environmental noises in the actual detection working environment will seriously reduce the signal-to-noise ratio of the cabinet response signal, thereby affecting the final accuracy of cabinet status detection.

[0045] In this embodiment, the transfer function method is used to perform noise reduction on the response signal, and the true spectrum of each measurement point is estimated. The calculation method is as follows

[0046]

[0047] in, For p i The cross spectrum between (f) and r(f), r(f) is the reference signal located near the acoustic excitation device, p i (f) is the signal of the remaining corner point measurement points, C rr (f) = E[r * (f)r(f)], is the autospectrum of r(f), E[·] represents the mathematical expectation, i = 1, 2, …, K-1, K is the total number of microphones.

[0048] S4: Calculate the cabinet acoustic clustering signal through the acoustic clustering algorithm and extract the cabinet acoustic cluster features.

[0049] The general detection characteristics of empty box detection (such as resonance frequency, response attenuation rate, etc.) have the defects of poor sensitivity and instability in practical applications, and are easily affected by interference such as environmental noise, local modes of the box wall panels, and differences in the box material.

[0050] In order to solve the above problems, in this embodiment, an acoustic clustering algorithm is first used to calculate the cabinet acoustic clustering signal. The calculation method is as follows:

[0051] The signal vectors of the four corner points on the surface of each box where the measuring points are arranged Use Hadamard matrix T to perform matrix transformation to obtain the cabinet acoustic clustering signal

[0052] Where T = (t1 t2 t3 t4), t1 = (1, 1, 1, 1), t2 = (1, 1, -1, -1), t3 = (1, -1, -1, 1), t4 = (1, -1, 1, -1).

[0053] Figure 3 The calculation results of acoustic clustering signals on the two measurement surfaces of the box are given respectively.

[0054] Then, the following method is used to extract the cabinet acoustic cluster features:

[0055] Acoustic clustering signal of the jth box surface where measurement points are arranged Assume that the spectrum values ​​at the analysis frequencies k-1, k and k+1 are and Set the threshold ratio δ, record like and Zeji The cluster feature is 1, otherwise it is 0. For example, set the threshold ratio δ = 5, Figure 3 The acoustic cluster characteristics of the two surfaces of the middle box at 17.5Hz are [1 0 0 0] and [0 1 0 0] respectively.

[0056] The above method obtains two acoustic cluster features S1(k) and S2(k) of the box surface with measurement points at frequency point k, and finally combines them into the acoustic cluster feature [S1(k), S2(k)] of the box. For example, Figure 3 The combined acoustic cluster feature of the middle box at 17.5Hz is [1 0 0 0 0 1 0 0].

[0057] S5: Compare the extracted acoustic cluster features with the reference acoustic cluster features of the cabinet identified in step S1 to achieve intelligent detection and judgment of the cabinet status.

[0058] In this embodiment, the intelligent detection and judgment method of the cabinet status is as follows:

[0059] The acoustic cluster characteristics of the inspected box [s1(k),…,s j (k)] is recorded as vector a, and step S1 identifies the reference acoustic cluster features of the box Denote it as vector b, and calculate the cosine of the angle between them cos(a,b) =<a,b> / (||a||||b||).

[0060] If the box state is the same as the empty box state identified in step S1, then cos(a,b)=1, so it can be judged as an empty box state, otherwise it is a non-empty box state.

[0061] For example, Figure 3The combined acoustic cluster feature a of the middle box at 17.5Hz is [1 0 0 0 0 1 0 0]. If the reference acoustic cluster feature b of the box identified in step S1 is [1 0 0 0 0 1 0 0], then cos(a,b)=1, which means the box is empty.

[0062] This embodiment also provides an acoustic intelligent detection system for empty containers, which mainly consists of a visual detection module, an acoustic excitation and receiving module, a support platform including an automatic deployment device, a data acquisition and preprocessing module, and a computer control system including relevant algorithms. The structure is shown in the figure below. Figure 4 shown.

[0063] Among them, the visual inspection module is used to detect the size and position information of the box; the acoustic excitation and receiving module is used to acoustically excite the box at a specified position and receive the acoustic response signal of the box at a specified position; the bracket platform including the automatic deployment device is used to install the visual inspection module, acoustic excitation and receiving module and other equipment, and automatically deploy and recover the acoustic excitation and receiving module according to requirements; the data acquisition and preprocessing module is used to collect signals such as the acoustic response of the box and complete preprocessing functions such as filtering; the computer control system is used to execute all visual inspection algorithms, automatic deployment and recovery control instructions, data acquisition and analysis processing algorithms, data reading and writing and result output and other programs and algorithms.

[0064] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for acoustic intelligent detection of empty containers, characterized by: The following steps are involved: S1: Use visual detection methods to obtain the size and position information of the entry box, and use the automatic placement device to accurately place the acoustic excitation and acoustic receiving devices at the corresponding designated positions; S2: The cabinet is acoustically excited by the acoustic excitation device, and the acoustic response signal of the cabinet is synchronously collected by the microphone in the acoustic receiving device; S3: Use the transfer function method to perform noise reduction on the response signal to obtain the true spectrum estimation of each measurement point; S4: Calculate the cabinet acoustic clustering signal through the acoustic clustering algorithm and extract the cabinet acoustic cluster features; S5: Compare the extracted acoustic cluster features with the reference acoustic cluster features of the cabinet identified in step S1 to achieve intelligent detection and judgment of the cabinet status; The acoustic cluster features of the box in S4 are extracted using the following method: For the first j Acoustic clustering signals on the surface of a box with measurement points arranged , set it at the analysis frequency k -1. k and k The spectral values ​​of the three points +1 are and , set the threshold ratio ,remember , ,like and , then remember The cluster feature is 1, otherwise it is 0, thus we get the j Frequency points on the cabinet surface k Acoustic cluster characteristics Finally, the acoustic cluster features of the box surface with all measurement points arranged are combined into the acoustic cluster features of the box .

2. The method for acoustic intelligent detection of empty containers according to claim 1, characterized in that: The number of the acoustic excitation device in S1 is 1, and its designated working position is located at a corner point of the box.

3. The method for acoustic intelligent detection of empty containers according to claim 2, characterized in that: The number of microphones in the sound receiving device in S1 is multiple, one of which is a designated working position at the same corner point of the box as the position of the sound excitation device, and the remaining designated working positions are located at the corner points of at least two box surfaces.

4. The method for acoustic intelligent detection of empty containers according to claim 1, characterized in that: The true spectrum estimation of each measurement point in S3 , the calculation method is as follows: ,in, ,for and The mutual spectrum between is the reference acoustic signal located near the acoustic excitation device, is the acoustic signal of the measuring points at the other corner positions, ,for 's autobiography, represents the mathematical expectation, i =1,2,…, K -1, K is the total number of microphones.

5. The method for acoustic intelligent detection of empty containers according to claim 1, characterized in that: The cabinet acoustic clustering algorithm in S4 is calculated as follows: The signal vectors of the four corner points on the surface of each box where the measuring points are arranged Using the Hadamard matrix Perform matrix transformation to obtain the cabinet acoustic clustering signal : ,in , , , , .

6. The method for acoustic intelligent detection of empty containers according to claim 4, characterized in that: The intelligent detection and judgment method of the cabinet status in S5 is as follows: Acoustic cluster characteristics of the inspected box Denoted as vector a, step S1 identifies the reference acoustic cluster features of the box Denote it as vector b and calculate the cosine of the angle between them ,like , it is judged to be an empty box state, otherwise it is a non-empty box state.

7. A system for implementing the method for acoustic intelligent detection of empty containers according to claim 1, characterized in that: It consists of a visual detection module, an acoustic excitation and receiving module, a support platform including an automatic deployment device, a data acquisition and preprocessing module, and a computer control system including relevant algorithms; The visual detection module is used to detect the size and position information of the entry box; The acoustic excitation and receiving module is used to acoustically excite the box at a specified position and receive an acoustic response signal of the box at the specified position; The support platform including the automatic deployment device is used to install the visual detection module, the acoustic excitation and receiving module equipment, and automatically deploy and recover the acoustic excitation and receiving modules as required; The data acquisition and preprocessing module is used to collect the cabinet acoustic response signal and complete the filtering preprocessing function; The computer control system is used to execute all visual inspection, automatic placement and recovery control, data collection and analysis processing, data reading and writing, and result output programs and algorithms.

Images