Target identification method based on multi-channel infrared camera

Abstract

The invention relates to the technical field of instrument detection equipment, in particular to a target identification method based on a multi-channel infrared camera. Comprising the following steps: S1, sequentially carrying out image acquisition on a target to be identified and a white board with 100% diffuse reflection by using a multi-channel infrared camera, and correspondingly obtaining n infrared images and n white board images; the multi-channel infrared camera is placed in a darkroom environment for shooting, dark noise of the multi-channel infrared camera is obtained, the reflectivity of the to-be-recognized target in the n-th spectral band is calculated, and a reflectivity matrix of the to-be-recognized target in the n-th spectral band is obtained; s2, normalizing the reflectivity of the to-be-identified target in the n spectral bands to obtain a first feature matrix of the to-be-identified target; s3, performing principal component analysis dimension reduction on the first feature matrix to obtain dimension-reduced data; and S4, identifying the target type of the to-be-identified target by using the data after dimension reduction. By means of the plastic sorting device, low-cost and high-efficiency sorting of plastic can be achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of instrument detection equipment, and in particular to a target recognition method based on a multi-channel infrared camera. Background Art

[0002] Plastics have been widely used in production and life due to their low cost, easy processing and strong durability. They are important basic materials. At present, my country's plastics industry is developing rapidly. The irregular recycling and disposal of a large amount of plastic waste has caused great energy waste and serious pollution to the environment. At present, there are many technologies for identifying plastic types, but the recognition accuracy is low, the test takes a long time, and the accuracy is not high, which seriously restricts the production capacity of plastic recycling companies. Existing methods for identifying plastic types using spectral signals include near-infrared hyperspectral imaging and single-probe scanning imaging.

[0003] Although the existing detection methods can basically meet the requirements of sorting, they have at least the following shortcomings: 1. High detection cost: The equipment of near-infrared hyperspectral imaging method is expensive, mainly using grating as the core spectroscopic original, and the components are imported and expensive. Near-infrared hyperspectral imaging requires the use of area array cameras. Therefore, the cost of setting up a waste plastic identification system is very high. 2. Low detection efficiency: Since the single-probe scanning imaging method requires point-by-point scanning of the entire identification area, it takes a long time and has a slow detection speed, which is especially not suitable for spectral collection of waste plastics in motion. Summary of the invention

[0004] In view of this, the present invention aims to provide a target recognition method based on a multi-channel infrared camera to solve the problems of high detection cost and low detection efficiency in the prior art.

[0005] To achieve the above object, the technical solution created by the present invention is implemented as follows: A target recognition method based on a multi-channel infrared camera specifically comprises the following steps: S1: Use a multi-channel infrared camera to collect images of the target to be identified and the 100% diffuse reflective white board in turn, and obtain n infrared images and n white board images accordingly; The multi-channel infrared camera is placed in a dark room environment for shooting, the dark noise of the multi-channel infrared camera is obtained, and the reflectivity of the target to be identified in the nth spectrum segment is calculated to obtain the reflectivity matrix of the target to be identified in the nth spectrum segment; S2: normalizing the reflectance of the target to be identified in n spectral bands to obtain the first feature matrix of the target to be identified; S3: Performing principal component analysis on the first feature matrix to reduce its dimension and obtain the data after dimension reduction; S4: Use the reduced-dimensional data to identify the target type.

[0006] Furthermore, in step S1, the multi-channel infrared camera sequentially acquires images of the target to be identified and the 100% diffuse reflective whiteboard, and shoots the images in a darkroom environment, with the same exposure time.

[0007] Furthermore, the number of line arrays of the multi-channel infrared camera is n.

[0008] Furthermore, in step S1, the reflectivity of the target to be identified in the nth spectral band is calculated. The calculation formula used is: (1); in, is the nth infrared image, is the nth whiteboard image, is the dark noise of the multi-channel infrared camera; Obtain the reflectivity matrix of the target to be identified in n spectral bands .

[0009] Furthermore, step S2 specifically includes the following steps: S21: Use the reflectance of the n spectral segments of the target to be identified to construct an n-dimensional vector to be identified, and calculate the similarity SAM between the n-dimensional vector to be identified and the n-dimensional calibration vector by the following formula: (2); Among them, X is the n-dimensional vector to be identified, and Y is the n-dimensional calibration vector; The ratio spectral coefficient matrix SR of the target to be identified is obtained by the following formula: (3); in, is the reflectivity of the target to be identified in the nth spectral band, is the reflectivity of the target to be identified in the n-1th spectral band, is the ratio spectral coefficient matrix of the target to be identified in the n-1th spectral segment, ; The n reflectances of the target to be identified are normalized by the following formula to obtain the normalized spectral coefficient matrix of the target to be identified: : (4); in, is the normalized spectral coefficient of the target to be identified in the n-1th spectral band, ; S22: Based on equations (2) to (4), construct the first feature matrix of the target to be identified .

[0010] Furthermore, in step S21, the target to be identified is replaced with a known target, and step S1 is repeated to obtain the reflectivity of n spectral segments of the known target, and an n-dimensional calibration vector is constructed using the reflectivity of n spectral segments of the known target.

[0011] Furthermore, step S3 specifically includes the following steps: S31: Obtain the first feature matrices of m types of targets, and organize the first feature matrices of the m types of targets into a feature matrix M with 3n-1 rows and m columns by columns: (5); in, is the first characteristic matrix of the first target, is the first characteristic matrix of the second target, is the first characteristic matrix of the mth target; S32: performing zero mean processing on each row element of the feature matrix M; S33: Calculate the covariance matrix C by the following formula: (6); S34: Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, and arrange the eigenvectors into a matrix in rows according to the corresponding eigenvalues ​​from small to large and from top to bottom, and take the first k rows of the matrix to form the matrix P, and obtain the data matrix F=PM reduced to k dimensions.

[0012] Furthermore, the data matrix .

[0013] Furthermore, step S4 specifically includes the following steps: S41: Based on the first feature matrix, obtain the second feature matrix of the target to be identified by the following formula : (7); S42: Obtain the third characteristic matrix of the i-th target ; S43: The Euclidean distances between the second feature matrix of the target to be identified and the third feature matrix of the i-th target are calculated in sequence by the following formula to obtain a set of Euclidean distances between the second feature matrix of the target to be identified and the third feature matrix of the m-th target: : (8); in, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the i-th target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the first target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the second target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the kth target; S44: The target category corresponding to the minimum Euclidean distance is used as the target category of the target to be identified.

[0014] A target recognition system based on a multi-channel infrared camera comprises a transmission mechanism, a multi-channel infrared camera, a host computer and a material separation mechanism. The multi-channel infrared camera is placed directly above the transmission mechanism, the multi-channel infrared camera is connected to the host computer, and the host computer is connected to the material separation mechanism.

[0015] Compared with the prior art, the invention can achieve the following beneficial effects: The target recognition method based on a multi-channel infrared camera created by the present invention utilizes a multi-channel hyperspectral infrared camera with a more compact structure and higher optical-mechanical integration to achieve low-cost and high-efficiency sorting of targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings constituting part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments and descriptions of the present invention are used to explain the present invention and do not constitute an improper limitation on the present invention. In the drawings: Figure 1 A schematic diagram of a flow chart of a target recognition method based on a multi-channel infrared camera according to an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a target recognition system based on a multi-channel infrared camera according to an embodiment of the present invention.

[0017] Description of reference numerals: 1. Conveying mechanism; 2. Multi-channel infrared camera; 3. Host computer; 4. Material dividing mechanism. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and do not constitute a limitation of the invention.

[0019] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0020] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions 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 a limitation on the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0021] In the description of the invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the invention can be understood according to specific circumstances.

[0022] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0023] like Figure 1 As shown, the target recognition method based on a multi-channel infrared camera provided by the present invention specifically includes the following steps: S1: Use a multi-channel infrared camera 2 to collect images of the target to be identified and the 100% diffuse reflective white board in sequence, and obtain n infrared images and n white board images accordingly; The multi-channel infrared camera 2 is placed in a dark room environment for shooting, the dark noise of the multi-channel infrared camera 2 is obtained, and the reflectivity of the target to be identified in the nth spectrum segment is calculated to obtain the reflectivity matrix of the target to be identified in the nth spectrum segment; In step S1, the multi-channel infrared camera 2 sequentially collects images of the target to be identified and the 100% diffuse reflective whiteboard, and shoots in a darkroom environment, with the same exposure time. The number of line arrays of the multi-channel infrared camera 2 is n.

[0024] In step S1, the reflectivity of the target to be identified in the nth spectrum segment is calculated The calculation formula used is: (1); in, is the nth infrared image, is the nth whiteboard image, is the dark noise of multi-channel infrared camera 2; Obtain the reflectivity matrix of the target to be identified in n spectral bands .

[0025] S2: normalizing the reflectance of the target to be identified in n spectral bands to obtain the first feature matrix of the target to be identified; S21: Use the reflectance of the n spectral segments of the target to be identified to construct an n-dimensional vector to be identified, and calculate the similarity SAM between the n-dimensional vector to be identified and the n-dimensional calibration vector by the following formula: (2); Among them, X is the n-dimensional vector to be identified, and Y is the n-dimensional calibration vector; The ratio spectral coefficient matrix SR of the target to be identified is obtained by the following formula: (3); in, is the reflectivity of the target to be identified in the nth spectral band, is the reflectivity of the target to be identified in the n-1th spectral band, is the ratio spectral coefficient matrix of the target to be identified in the n-1th spectral segment, ; The n reflectances of the target to be identified are normalized by the following formula to obtain the normalized spectral coefficient matrix of the target to be identified: : (4); in, is the normalized spectral coefficient of the target to be identified in the n-1th spectral band, ; In step S21, the target to be identified is replaced with a known target, and step S1 is repeated to obtain the reflectivity of n spectral segments of the known target, and an n-dimensional calibration vector is constructed using the reflectivity of n spectral segments of the known target.

[0026] S22: Based on equations (2) to (4), construct the first feature matrix of the target to be identified .

[0027] S3: Performing principal component analysis on the first feature matrix to reduce its dimension and obtain the data after dimension reduction; S31: Obtain the first feature matrices of m types of targets, and organize the first feature matrices of the m types of targets into a feature matrix M with 3n-1 rows and m columns by columns: (5); in, is the first characteristic matrix of the first target, is the first characteristic matrix of the second target, is the first characteristic matrix of the mth target; S32: performing zero mean processing on each row element of the feature matrix M; S33: Calculate the covariance matrix C by the following formula: (6); S34: Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, and arrange the eigenvectors into a matrix in rows according to the corresponding eigenvalues ​​from small to large and from top to bottom, and take the first k rows of the matrix to form the matrix P, and obtain the data matrix F=PM reduced to k dimensions. Data matrix .

[0028] Solving the eigenvalues ​​of the covariance matrix and the corresponding eigenvectors is as follows: v is the eigenvector of the covariance matrix, is the eigenvalue of the covariance matrix, I is the identity matrix, and for the characteristic polynomial equation , find all possible Values, these values ​​are the eigenvalues ​​of the matrix M; for each eigenvalue obtained , substituting it into the characteristic polynomial equation above can solve v and obtain the eigenvector corresponding to the characteristics of the covariance matrix.

[0029] S4: Use the reduced-dimensional data to identify the target type.

[0030] S41: Based on the first feature matrix, obtain the second feature matrix of the target to be identified by the following formula : (7); S42: Obtain the third characteristic matrix of the i-th target ; S43: The Euclidean distances between the second feature matrix of the target to be identified and the third feature matrix of the i-th target are calculated in sequence by the following formula to obtain a set of Euclidean distances between the second feature matrix of the target to be identified and the third feature matrix of the m-th target: : (8); in, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the i-th target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the first target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the second target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the kth target; S44: The target category corresponding to the minimum Euclidean distance is used as the target category of the target to be identified.

[0031] The target recognition method based on a multi-channel infrared camera needs to be combined with a target recognition system based on a multi-channel infrared camera to achieve automatic sorting of target types. Taking plastic sorting as an example, the structure of the target recognition system based on a multi-channel infrared camera is introduced below. Figure 2 As shown, the target recognition system based on the multi-channel infrared camera includes a conveying mechanism 1, a multi-channel infrared camera 2, a host computer 3 and a material dividing mechanism 4. The multi-channel infrared camera 2 is placed directly above the conveying mechanism 1, the multi-channel infrared camera 2 is connected to the host computer 3, and the host computer 3 is connected to the material dividing mechanism 4; The conveying mechanism 1 transports the plastic to be identified, and the multi-channel infrared camera 2 collects the image of the plastic to be identified; the upper computer 3 determines the type of the plastic to be identified based on the image of the plastic to be identified, and converts the determination result into the control parameter of the sorting mechanism 4. The sorting mechanism 4 separates the plastic to be identified according to the control parameter, so that the plastic to be identified falls into the corresponding collection device.

[0032] The number of collecting devices is the same as the number of plastic types that can be identified by the multi-channel infrared camera 2. The position of the collecting devices is arranged accordingly according to the jet direction of the sorting mechanism 4 and the conveying direction of the conveyor belt of the conveying mechanism 1. For example, the installation position of the collecting device is set to correspond to the jet direction of the sorting mechanism 4, and the plastic blown off from the conveying mechanism 1 by the sorting mechanism 4 is collected, etc. The sorting mechanism 4 includes an air valve, and the target sorting method can be replaced from the air valve to a robot arm and a suction nozzle. The conveying mechanism 1 uses an 800mm wide V-shaped conveyor belt (manufacturer: Chaohua Rubber Conveyor Belt Manufacturing Co., Ltd.).

[0033] The target recognition method based on a multi-channel infrared camera created by the present invention is not limited to the plastic sorting industry. In the field of nut sorting, after the nuts are crushed, the flesh and the shell are mixed together. The traditional nut sorting method relies on manual selection, and the efficiency and accuracy cannot be guaranteed. The method proposed by the present invention can greatly improve its production efficiency. In agricultural applications, the method proposed by the present invention can be used to detect and identify grains and spoiled grains, and to determine the germination of seeds. In the field of medicine, the method proposed by the present invention can be used to detect whether there are cracks in the sustained-release coating of tablets, whether there are tablets in capsules or blisters, the spectral differences of drugs of the same color but different ingredients, and to detect and analyze the sealing of the left and right parts of the capsule.

[0034] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0035] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A target recognition method based on a multi-channel infrared camera, characterized in that: The specific steps include: S1: using the multi-channel infrared camera to sequentially capture images of the target to be identified and the 100% diffuse reflective whiteboard, and correspondingly obtaining n infrared images and n whiteboard images; The multi-channel infrared camera is placed in a dark room environment for shooting, dark noise of the multi-channel infrared camera is obtained, and the reflectivity of the target to be identified in the nth spectrum segment is calculated to obtain a reflectivity matrix of the target to be identified in the nth spectrum segment; S2: normalizing the reflectance of the target to be identified in n spectral bands to obtain a first characteristic matrix of the target to be identified; S3: Performing principal component analysis on the first feature matrix to reduce its dimension, and obtaining data after dimension reduction; S4: Using the data after dimensionality reduction to identify the target type of the target to be identified.

2. The target recognition method based on a multi-channel infrared camera according to claim 1, characterized in that: In step S1, the multi-channel infrared camera sequentially collects images of the target to be identified and the 100% diffuse reflective white board, and shoots them in a darkroom environment, with the same exposure time.

3. The target recognition method based on a multi-channel infrared camera according to claim 1, characterized in that: The number of line arrays of the multi-channel infrared camera is n.

4. The target recognition method based on a multi-channel infrared camera according to claim 1, characterized in that: In step S1, the reflectivity of the target to be identified in the nth spectrum segment is calculated. The calculation formula used is: （1）； in, is the nth infrared image, is the nth whiteboard image, is the dark noise of the multi-channel infrared camera; Obtain the reflectivity matrix of the target to be identified in n spectral bands .

5. The target recognition method based on a multi-channel infrared camera according to claim 1, characterized in that: The step S2 specifically includes the following steps: S21: construct an n-dimensional vector to be identified using the reflectance of the n spectral segments of the target to be identified, and calculate the similarity SAM between the n-dimensional vector to be identified and the n-dimensional calibration vector by the following formula: （2）； Among them, X is the n-dimensional vector to be identified, and Y is the n-dimensional calibration vector; The ratio spectrum coefficient matrix SR of the target to be identified is obtained by the following formula: （3）； in, is the reflectivity of the target to be identified in the nth spectral band, is the reflectivity of the target to be identified in the n-1th spectral band, is the ratio spectral coefficient matrix of the target to be identified in the n-1th spectral segment, ; The normalized spectral coefficient matrix of the target to be identified is obtained by normalizing the n reflectances of the target to be identified using the following formula: : （4）； in, is the normalized spectral coefficient of the target to be identified in the n-1th spectral band, ; S22: Based on equations (2) to (4), construct a first feature matrix of the target to be identified .

6. The target recognition method based on a multi-channel infrared camera according to claim 5 is characterized in that: In step S21, the target to be identified is replaced with a known target, and step S1 is repeated to obtain the reflectivity of n spectral segments of the known target, and an n-dimensional calibration vector is constructed using the reflectivity of n spectral segments of the known target.

7. The target recognition method based on a multi-channel infrared camera according to claim 6 is characterized in that: The step S3 specifically includes the following steps: S31: Obtain the first feature matrices of m types of targets, and organize the first feature matrices of the m types of targets into a feature matrix M with 3n-1 rows and m columns by columns: (5)； in, is the first characteristic matrix of the first target, is the first characteristic matrix of the second target, is the first characteristic matrix of the mth target; S32: performing zero mean processing on each row element of the feature matrix M; S33: Calculate the covariance matrix C by the following formula: (6)； S34: Solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, and arrange the eigenvectors into a matrix in rows according to the corresponding eigenvalues ​​from small to large and from top to bottom, and take the first k rows of the matrix to form a matrix P, and obtain a data matrix F=PM reduced to k dimensions.

8. The target recognition method based on a multi-channel infrared camera according to claim 7 is characterized in that: Data Matrix .

9. The target recognition method based on a multi-channel infrared camera according to claim 8, characterized in that: The step S4 specifically includes the following steps: S41: Based on the first feature matrix, obtain the second feature matrix of the target to be identified by the following formula : （7）； S42: Obtain the third characteristic matrix of the i-th target ; S43: The Euclidean distances between the second feature matrix of the target to be identified and the third feature matrix of the i-th target are calculated in sequence by the following formula to obtain a set of Euclidean distances between the second feature matrix of the target to be identified and the third feature matrix of the m-th target: : （8）； in, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the i-th target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the first target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the second target, is the Euclidean distance between the second feature matrix of the target to be identified and the third feature matrix of the kth target; S44: The target category corresponding to the minimum Euclidean distance is used as the target category of the target to be identified.

10. A target recognition system based on a multi-channel infrared camera, characterized in that: It includes a conveying mechanism, a multi-channel infrared camera, a host computer and a material dividing mechanism. The multi-channel infrared camera is placed directly above the conveying mechanism. The multi-channel infrared camera is connected to the host computer, and the host computer is connected to the material dividing mechanism. The target recognition system based on the multi-channel infrared camera is used to execute the target recognition method based on the multi-channel infrared camera described in any one of claims 1 to 9.

Images