Multivariate fusion information early warning detection method for contraband inside parcel
By collecting and fusing information from X-ray images, odor and microwave signals, a contraband detection model is constructed, which solves the problem of missed detection of X-ray security machine detection in the existing technology, and achieves more efficient and accurate contraband detection .
Patent Information
- Application Number
- CN202410941216.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art relies on X-ray security machine to detect the X-ray image of luggage incorrectly and missed, and the long-term low concentration of security personnel leads to fatigue, making it difficult to effectively detect the diversity of contraband.
Collect X-ray image information, odor information and microwave signals, perform pre-processing and feature extraction, and integrate features to build a contraband detection model to achieve automated detection.
Through the multi-information fusion detection method, the accuracy and efficiency of contraband detection are improved, the false detection caused by fatigue by security personnel is reduced, and the contraband can be identified more comprehensively.
Smart Images

Figure CN118941766A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of security monitoring, and in particular relates to a multi-information fusion early warning detection method for contraband in a package. Background Art
[0002] At present, the detection of prohibited items in luggage mainly relies on X-ray security inspection machines to obtain X-ray images of luggage, and then security inspectors observe the X-ray images and make judgments. This process takes only a few seconds; the uneven professional level of security inspectors and visual fatigue caused by long-term focused observation will lead to false detection and missed detection of prohibited items. The long-term boring and monotonous work of security inspectors will increase work inertia and generate additional work pressure.
[0003] Moreover, X-ray image detection can only detect shapes, which is not enough to cope with the diversity of contraband and is likely to miss contraband. Summary of the invention
[0004] In order to solve the above technical problems, the present invention proposes a multi-information fusion early warning detection method for prohibited items in a package to solve the problems existing in the above-mentioned prior art.
[0005] To achieve the above object, the present invention provides a multi-information fusion early warning detection method for contraband in a package, comprising the following steps:
[0006] Collect X-ray image information, odor information and microwave signals of contraband;
[0007] Preprocessing the X-ray image information, odor information and microwave signal to obtain a preprocessed signal;
[0008] Extracting features from the preprocessed signal to obtain multivariate features, and fusing the multivariate features to obtain fused features;
[0009] Constructing an initial contraband detection model and training it through the fusion features to obtain a contraband detection model;
[0010] The package is inspected by the contraband detection model to obtain a detection result.
[0011] Preferably, the method for preprocessing the X-ray image information and odor information comprises:
[0012] Performing filtering processing on the X-ray image information and the sensor array response image representing the odor information to remove image noise;
[0013] Data enhancement processing is performed based on the X-ray image information and the sensor array response image representing the odor information. The data enhancement processing method includes acquiring a new image by random rotation and image flipping.
[0014] Preferably, the method for preprocessing the microwave signal comprises:
[0015] The microwave signal is filtered by using a T1 wavelet filtering algorithm.
[0016] Preferably, the process of extracting features from the preprocessed X-ray image information includes:
[0017] Use the Sobel operator to calculate pixel gradients;
[0018] The non-maximum pixel gradient suppression method is used to eliminate the stray effect;
[0019] Eliminate small gradient modulus points through threshold hysteresis processing;
[0020] By isolating weak edges and suppressing them, edge points are screened out to obtain the contraband edge point set;
[0021] The shape features of contraband are obtained based on the edge point set of contraband.
[0022] Preferably, the process of extracting features from the preprocessed odor information includes:
[0023] By fitting the response curve of the sensor array response image, the amplitude and time constant are extracted as features.
[0024] Preferably, the process of extracting features from the preprocessed microwave signal includes:
[0025] Select wavelet basis functions according to microwave signal characteristics;
[0026] Determine the decomposition level of wavelet transform based on the frequency content of microwave signal and the required time-frequency resolution;
[0027] Performing multi-scale decomposition on the preprocessed microwave signal according to the wavelet basis function and the decomposition layer to obtain components on different time scales;
[0028] Wavelet transform coefficients are obtained based on components on different time scales, and the characteristics of microwave signals are obtained based on the wavelet transform coefficients.
[0029] Preferably, the process of obtaining the contraband detection model includes:
[0030] An initial contraband detection model is constructed based on a convolutional neural network, and the initial contraband detection model is trained and verified through the fusion features to obtain a contraband detection model.
[0031] Preferably, the process of obtaining the contraband detection model further includes:
[0032] Constructing an SE module, and adjusting the weight learning process of the SE module in a timely manner through adaptive adaptation to obtain an SE module with generalization ability;
[0033] The SE module with generalization ability is inserted after the activation function of each convolution of the convolutional neural network.
[0034] Compared with the prior art, the present invention has the following advantages and technical effects:
[0035] The present invention discloses a multivariate fusion information early warning detection method for prohibited items in a package, comprising the following steps: collecting X-ray image information, odor information and microwave signals of prohibited items; preprocessing the X-ray image information, odor information and microwave signals to obtain preprocessed signals; extracting features from the preprocessed signals to obtain multivariate features, fusing the multivariate features to obtain fused features; constructing an initial prohibited item detection model and training it through fused features to obtain a prohibited item detection model; and detecting the package through the prohibited item detection model to obtain a detection result. The present invention can obtain multivariate features through a variety of collection methods, can avoid missed detection to the greatest extent, and can avoid false detection caused by fatigue of security personnel through automatic detection through the prohibited item detection model. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0037] Figure 1 This is a flow chart of a method for early warning detection of contraband in a package according to an embodiment of the present invention using multi-information fusion. DETAILED DESCRIPTION
[0038] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0039] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0040] Embodiment 1
[0041] like Figure 1 As shown, this embodiment provides a multi-information fusion early warning detection method for prohibited items in a package, comprising the following steps:
[0042] Collect X-ray image information, odor information and microwave signals of contraband;
[0043] Preprocessing the X-ray image information, odor information and microwave signal to obtain a preprocessed signal;
[0044] Extracting features from the preprocessed signal to obtain multivariate features, and fusing the multivariate features to obtain fused features;
[0045] Constructing an initial contraband detection model and training it through the fusion features to obtain a contraband detection model;
[0046] The package is inspected by the contraband detection model to obtain a detection result.
[0047] This embodiment uses a variety of technical means such as X-ray image information, odor information and microwave signals to detect dangerous objects such as dangerous tools, dangerous chemicals and heavy metals, thereby improving the accuracy and efficiency of contraband detection. Below, we will elaborate on the various steps and principles of this method:
[0048] First, the method uses professional equipment to collect X-ray image information, odor information and microwave signals of the package. X-ray image information can intuitively show the structure of the items inside the package, while odor information helps to identify the unique odor characteristics of certain specific contraband. Microwave signals can reflect the electromagnetic properties of the materials inside the package, providing important clues for detection.
[0049] After collecting this information, the next step is to preprocess it. For X-ray image information and odor information, filtering is usually used to remove image noise and improve the clarity of information. In addition, data enhancement processing, such as random rotation and image flipping, is used to generate more training samples and improve the generalization ability of the model. For microwave signals, the T1 wavelet filtering algorithm is used to filter to eliminate interference components in the signal.
[0050] Filtering includes Gaussian filtering:
[0051] The purpose of Gaussian filtering is to filter out Gaussian noise in the image. The principle is to use the isotropy and separability of the Gaussian kernel to create a suitable Gaussian kernel to perform weighted averaging on the pixel values in the image. Since the image is two-dimensional, it belongs to a two-dimensional Gaussian distribution, and the following equation can be listed:
[0052]
[0053] Considering that the image pixels exist in a discrete form in the matrix, discrete approximation is required. The Gaussian kernel set here is [2,4,5,4,2; 4,9,12,9,4; 5,12,15,12,5; 4,9,12,9,4; 2,4,5,4,2]. In actual operation, the convolution matrix is used to scan each pixel in the waterline image, and the weighted average of the pixels in the neighborhood determined by the convolution is used to replace the value of the convolution center pixel.
[0054] After preprocessing, it is necessary to extract features from these signals. For X-ray image information, the Sobel operator can be used to calculate pixel gradients, and the non-maximum pixel gradient suppression method can be used to eliminate stray effects. Then, through threshold hysteresis processing and isolated weak edge suppression, the edge point set of contraband can be screened out, and then the shape characteristics of the contraband can be obtained. For odor information, the response curve of the sensor array response image can be fitted to extract features such as amplitude and time constant. For microwave signals, the appropriate wavelet basis function is selected according to the characteristics of the microwave signal, and the components on different time scales are obtained through multi-scale decomposition, and features are extracted based on these components.
[0055] After extracting the multivariate features, they need to be fused to obtain more comprehensive information. The method of fusion features can be selected according to the actual situation, such as weighted average, maximum and minimum value fusion, etc. By fusing features, the complementarity of various information sources can be fully utilized to improve the accuracy of detection.
[0056] Next, an initial contraband detection model is constructed based on the convolutional neural network. Through the training process, the model is able to learn and identify the contraband patterns in the fused features. In order to improve the performance of the model, the SE module (Squeeze-and-Excitation module) can also be introduced to improve the generalization ability of the model by adaptively adjusting the weight learning process. The SE module can be inserted after each convolutional layer of the convolutional neural network to further improve the detection accuracy by enhancing the network's attention to important features.
[0057] After training and verification, the final prohibited goods detection model is obtained. The model can quickly and accurately detect packages and output the detection results. In practical applications, by placing the package on the detection equipment, the relevant information can be automatically collected and input into the detection model to achieve automated and intelligent prohibited goods detection.
[0058] In summary, this multi-information fusion early warning detection method for prohibited items in packages uses a variety of technical means to improve the accuracy and efficiency of prohibited items detection through the collection, preprocessing, feature extraction and fusion of multiple information. This method has broad application prospects in the fields of logistics security, customs supervision, etc., and is of great significance for ensuring public safety and maintaining social stability.
[0059] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A multi-information fusion early warning detection method for contraband in a package, characterized in that: The following steps are involved: Collect X-ray image information, odor information and microwave signals of contraband; Preprocessing the X-ray image information, odor information and microwave signal to obtain a preprocessed signal; Extracting features from the preprocessed signal to obtain multivariate features, and fusing the multivariate features to obtain fused features; Constructing an initial contraband detection model and training it through the fusion features to obtain a contraband detection model; The package is inspected by the contraband detection model to obtain a detection result.
2. The multi-information fusion early warning detection method for prohibited items in a package according to claim 1 is characterized in that: The method for preprocessing the X-ray image information and the odor information includes: Performing filtering processing on the X-ray image information and the sensor array response image representing the odor information to remove image noise; Data enhancement processing is performed based on the X-ray image information and the sensor array response image representing the odor information. The data enhancement processing method includes acquiring a new image by random rotation and image flipping.
3. The multi-information fusion early warning detection method for prohibited items in a package according to claim 1 is characterized in that: The method for preprocessing the microwave signal comprises: The microwave signal is filtered by using a T1 wavelet filtering algorithm.
4. The multi-information fusion early warning detection method for prohibited items in a package according to claim 1 is characterized in that: The process of feature extraction of preprocessed X-ray image information includes: Use the Sobel operator to calculate pixel gradients; The non-maximum pixel gradient suppression method is used to eliminate the stray effect; Eliminate small gradient modulus points through threshold hysteresis processing; By isolating weak edges and suppressing them, edge points are screened out to obtain the contraband edge point set; The shape features of contraband are obtained based on the edge point set of contraband.
5. The multi-information fusion early warning detection method for prohibited items in a package according to claim 1 is characterized in that: The process of feature extraction of preprocessed odor information includes: By fitting the response curve of the sensor array response image, the amplitude and time constant are extracted as features.
6. The multi-information fusion early warning detection method for prohibited items in a package according to claim 1 is characterized in that: The process of feature extraction of preprocessed microwave signals includes: Select wavelet basis functions according to microwave signal characteristics; Determine the decomposition level of wavelet transform based on the frequency content of microwave signal and the required time-frequency resolution; Performing multi-scale decomposition on the preprocessed microwave signal according to the wavelet basis function and the decomposition layer to obtain components on different time scales; Wavelet transform coefficients are obtained based on components on different time scales, and the characteristics of microwave signals are obtained based on the wavelet transform coefficients.
7. The multi-information fusion early warning detection method for prohibited items in a package according to claim 1 is characterized in that: The process of obtaining the contraband detection model includes: An initial contraband detection model is constructed based on a convolutional neural network, and the initial contraband detection model is trained and verified through the fusion features to obtain a contraband detection model.
8. The multi-information fusion early warning detection method for prohibited items in a package according to claim 7 is characterized in that: The process of obtaining the contraband detection model also includes: Constructing an SE module, and adjusting the weight learning process of the SE module in a timely manner through adaptive adaptation to obtain an SE module with generalization ability; The SE module with generalization ability is inserted after the activation function of each convolution of the convolutional neural network.
Citation Information
Cited By
Intelligent multi-mode classification detection system and method
CN120296472A
Risk early warning method and system for contraband detection equipment by using cloud network side
CN121746745A
A contraband detection device risk early warning method and system using a cloud network end
CN121746745B
