WO2022121290A1 - 一种基于果纹图谱和区块链的果品可信追溯方法及装置 - Google Patents

一种基于果纹图谱和区块链的果品可信追溯方法及装置 Download PDF

Info

Publication number
WO2022121290A1
WO2022121290A1 PCT/CN2021/103363 CN2021103363W WO2022121290A1 WO 2022121290 A1 WO2022121290 A1 WO 2022121290A1 CN 2021103363 W CN2021103363 W CN 2021103363W WO 2022121290 A1 WO2022121290 A1 WO 2022121290A1
Authority
WO
WIPO (PCT)
Prior art keywords
fruit
feature
hilum
image
fruit pattern
Prior art date
Application number
PCT/CN2021/103363
Other languages
English (en)
French (fr)
Inventor
钱建平
吴文斌
余强毅
史云
Original Assignee
中国农业科学院农业资源与农业区划研究所
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国农业科学院农业资源与农业区划研究所 filed Critical 中国农业科学院农业资源与农业区划研究所
Priority to US17/698,142 priority Critical patent/US20220207789A1/en
Publication of WO2022121290A1 publication Critical patent/WO2022121290A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation
    • G06T11/001—Texturing; Colouring; Generation of texture or colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/50—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols using hash chains, e.g. blockchains or hash trees
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/32—Normalisation of the pattern dimensions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/68—Food, e.g. fruit or vegetables
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/80—Recognising image objects characterised by unique random patterns
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/06—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols the encryption apparatus using shift registers or memories for block-wise or stream coding, e.g. DES systems or RC4; Hash functions; Pseudorandom sequence generators
    • H04L9/0643—Hash functions, e.g. MD5, SHA, HMAC or f9 MAC
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/32—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials
    • H04L9/3247—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials involving digital signatures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30128—Food products

Definitions

  • the invention relates to plant feature extraction, identification and block chain technology, in particular to information coding, identification and block chain traceability technology based on natural texture of fruits.
  • Invention patent - coding method and coding system for vegetable traceability discloses a coding system for vegetable traceability; also discloses a corresponding coding method for vegetable traceability .
  • Invention patent-method and device for tracing the source of animal individuals application number/patent number: 200910199977.X
  • application number/patent number: 200910199977.X application number/patent number: 200910199977.X
  • the converted animal identification code number is converted into the animal identification code number
  • the converted animal identification code number is stored in the electronic identification
  • the international animal code number in the electronic identification is read out for accessing international and domestic animal information data.
  • Invention patent - aquaculture product quality and safety management and traceability method and system provides a aquaculture product quality and safety management and traceability method and system, which is produced by solidifying pollution-free aquaculture standards
  • the database adopts the HACCP system for hazard analysis and critical control points, and provides a global unified code ( EAN/UCC) aquaculture management and traceability solutions supported by barcode automatic identification technology.
  • EAN/UCC global unified code
  • the first object of the present invention is to propose a method for encoding fruit pattern atlas information, which is used to generate a fruit pattern atlas, which can utilize the fruit pattern features of the fruit itself to construct an encoding method for a pattern based on the fruit pattern features, Establish an "identity card" for each fruit;
  • the second object of the present invention is to provide an identification method based on a fruit pattern, by comparing the fruit pattern of the fruit to be inspected on the client with the fruit pattern stored on the server, thereby realizing the single fruit identification of the fruit;
  • the third object of the present invention is to provide a credible traceability method for fruit products based on the fruit pattern and blockchain, which utilizes the blockchain technology to store the fruit pattern on the blockchain, so as to realize the retrieval of the fruit pattern data. cannot be tampered with;
  • the fourth object of the present invention is to provide a fruit pattern atlas generation and winding device for the above method
  • the purpose of the present invention is also to provide a system for credible traceability of fruit products based on fruit pattern map and blockchain.
  • a method for encoding fruit pattern map information characterized in that the method comprises the steps:
  • the fruit pattern map may further include a feature coding table of the pedicle part and/or the feature coding table of the hilum part.
  • One of the characteristics of the present invention is that the natural texture of the fruit is used as a reference, and the image features are extracted from the natural texture and encoded.
  • the image of the umbilical portion of the fruit has sufficient resolution to encode all the fruits in the world with different features, so the technical effect of the present invention lies in accuracy and uniqueness.
  • the present invention has the outstanding advantages of being anti-counterfeiting and facilitating traceability. It is especially suitable for the coding and identification of fruits with obvious lines, such as watermelon, cantaloupe, apple and other fruits.
  • one of the features of the present invention is that the feature of the rectangular image of the fruit pedicle and the rectangular image of the fruit hilum is used respectively, and the image feature coding is carried out to obtain the feature coding table of the pedicle and the feature coding table of the hilum.
  • the pedicle part feature code table and the fruit hilum part feature code table are combined to obtain the combined bidirectional feature code table. Therefore, when the feature code table matching is performed to identify fruits, the combined bidirectional feature code table can be used for matching first.
  • the combined bidirectional feature coding table can match, it means that the fruit to be identified completely matches the fruit stored in the database and can be used for fast matching; however, if the partial skin of the fruit is scratched, such as the partial rectangular image of the fruit pedicle and the When the shape or pattern of one of the rectangular images of the umbilicus has changed, so that the combined bidirectional feature coding table cannot be matched, the feature coding table of the pedicel and the umbilical part can be used for identification respectively. When only one feature encoding table can match and meet other conditions, it can also be regarded as successful identification. Therefore, the encoding and identification method of the present invention also has certain error tolerance, and can be compatible with detection efficiency and robustness.
  • the step of performing normalization processing on the image and converting it into a rectangular image of the pedicel part and the rectangular image of the hilum part includes:
  • I(x, y) represents the ring image
  • when r 1, it means I(x(r, ⁇ ), y(r, ⁇ ) ) is the pixel point of the outer edge of the ring image
  • histogram equalization is further adopted to enhance the image to obtain a clearer texture; the histogram equalization is as follows:
  • N is the total number of pixels in the image to be enhanced
  • N(r k ) is the number of pixels whose gray level is rk in the image
  • k is the number of gray levels
  • T(r k ) is the number of pixels for the gray level rk Conversion function
  • S(r k ) is the converted gray level.
  • the described extracting features of the rectangular image of the pedicel part and the rectangular image of the hilum part respectively, carrying out the image feature coding, and obtaining the feature coding table of the pedicel part and the umbilical part of the feature coding table comprises:
  • x 1 x cos ⁇ +y sin ⁇
  • y 1 -x sin ⁇ +y cos ⁇
  • wavelength ⁇ is specified in pixels
  • ⁇ represents the wavelength parameter of the cosine function in the Gabor kernel function, specified in pixels, usually greater than or equal to 2, but cannot be larger than one-fifth of the input image size
  • ⁇ is the direction; specifies the direction of the parallel stripes of the Gabor function, and its value is 0 to 360 degrees.
  • ⁇ represents the standard deviation of the Gaussian factor of the Gabor function.
  • h Re , h Im represent the real and imaginary parts of the filtered feature parameters, respectively, fft represents the Fourier transform, and ifft represents the inverse Fourier transform.
  • G(f) represents the filter, corresponding to the aforementioned G(x, y); I(r) represents the normalized image.
  • the combined operation is performed for the feature encoding table of the pedicle part and the feature encoding table of the hilum part, and the combined bidirectional feature encoding table is obtained as:
  • T is the combined bidirectional feature coding table
  • T A and TB are the feature coding tables of the pedicel and hilum, respectively
  • X(i,j) is the value corresponding to the i-th row and the j-th column in the bidirectional coding table .
  • the present invention provides a kind of identification method based on fruit pattern atlas, the method adopts the above-mentioned method to obtain the fruit pattern atlas, and step D, the fruit to be identified carries out image feature encoding, and the obtained
  • the feature encoding table is matched with the feature encoding table of the fruit pattern map, and is used for identifying the feature information of the fruit pattern map.
  • the described fruit to be identified carries out image feature encoding, and the obtained feature encoding table is matched with the stored feature encoding table, and the identification of the feature information of the fruit pattern map includes the steps:
  • N is the number of bits of feature encoding
  • XOR represents XOR operation
  • P j and Q j respectively represent the jth bit of texture feature encoding P and Q;
  • the distance is greater than the predetermined first threshold, it belongs to different fruits, and if it is determined that the distance is less than the predetermined first threshold, it belongs to the same fruit.
  • the described fruit to be identified carries out image feature encoding, and the obtained feature encoding table is matched with the stored feature encoding table, and the identification of the feature information of the fruit pattern map includes the steps:
  • the bidirectional feature encoding table Extract the feature coding table of the pedicle part and the feature coding table of the hilum part respectively, and further identify the feature information of the fruit pattern according to the feature coding table of the pedicel part and the umbilical part respectively.
  • One of the determined distances obtained from the table or the umbilical part feature encoding table is smaller than the predetermined first threshold, then it is only necessary to determine that the other of the distances is smaller than the second predetermined threshold to determine that it belongs to the same fruit and end the identification.
  • the present invention provides a kind of fruit credible traceability method based on fruit pattern atlas and block chain, which comprises adopting the above-mentioned method to obtain the fruit pattern atlas of fruit, and steps:
  • E. Blockchain on-chain The fruit pattern map is hashed and signed, and then uploaded to the chain to store the certificate, and the block number of the blockchain where the fruit pattern map is located is obtained;
  • the step E includes:
  • Pre-processing on the chain perform a hashing process on the fruit pattern map to obtain a series of hash values bound to the fruit pattern map, use an asymmetric encryption algorithm to sign the hash value, and confirm with the public key and signature information
  • the sender holds the corresponding private key, thereby converting the fruit pattern into a way readable by the blockchain, and at the same time binding the sender's identity to the sent information through a signature;
  • On-chain processing The processed data is sent to the blockchain node to form a blockchain transaction and enter the on-chain stage; after receiving the transaction, each node of the blockchain first broadcasts the received transaction to other The nodes form a unified transaction pool; the transaction data includes the hash of the packaged transactions in the block, and the transactions need to be sorted according to a unified order; after confirming the block header and hash, calculate the block hash, and pass the previous block.
  • the hash and its own hash are connected to form a chain to complete the chaining process.
  • the method for credible traceability of fruit products based on the fruit pattern map and the blockchain also includes the steps of hashing the traceability information of the fruit products, and then storing the hash values on the chain one by one to obtain the corresponding block number.
  • the block number and the traceability information are updated to the local application server in a one-to-one correspondence.
  • the traceability information includes fruit variety information, planting information, cold chain logistics information, and the like.
  • the method of uploading traceability information is the same as above.
  • the credible traceability method for fruit products based on the fruit pattern map and the blockchain also includes the steps:
  • Retrospective query use the above-mentioned method to obtain the fruit pattern spectrum of the fruit to be inspected, and use the described identification method to match the corresponding fruit pattern spectrum on the local application server through the fruit pattern spectrum of the fruit to be inspected, and obtain the area corresponding to the fruit.
  • Block number obtain the hash value of the fruit texture map stored on the blockchain through the block number, perform hash processing on the fruit texture map of the fruit stored on the local application server to obtain the local fruit texture map hash value, Consistently compare the hash value of the fruit pattern on the blockchain with the hash value of the local fruit pattern.
  • the traceability query adopts the above method to obtain the fruit pattern map of the fruit to be inspected, and obtains the traceability information of the fruit, the block number of the fruit pattern map and the block number of the traceability information on the local application server through the fruit pattern map, and then the block number of the fruit pattern to be inspected is obtained.
  • the fruit pattern map and traceability information of the fruit are hashed separately, and the consistency comparison with the hash value stored on the blockchain obtained through the block number is carried out to determine whether the traced fruit is the fruit corresponding to the fruit pattern map. And determine whether the traceability information has been tampered with. .
  • Consumers can use a device with a camera function, such as a smart phone or other smart devices, to trace the source by taking images of the stem portion and/or the hilum portion of the fruit using the above-mentioned traceability method.
  • a device with a camera function such as a smart phone or other smart devices
  • the present invention provides a fruit pattern atlas generation and on-chain device for the above method, the device comprises a fruit pattern atlas generation module and a blockchain on-chain module, wherein,
  • the fruit pattern atlas generation module includes an image acquisition device and a fruit pattern atlas feature code generation device, and the image acquisition device is used to capture images of the stem portion and the hilum portion of each fruit to be encoded; the atlas feature encoding
  • the generating device is used to grayscale the image of the pedicel part and the part of the hilum, and normalize the images to convert them into a rectangular image of the pedicel and the rectangular image of the hilum; and is used to extract the pedicel respectively Part of the rectangular image and part of the fruit hilum part of the rectangular image features, image feature coding, to obtain the pedicle part of the feature coding table and the fruit hilum part of the feature coding table, for the fruit pedicle part of the feature coding table and the fruit hilum part of the feature coding table, perform a parallel operation to obtain The combined bidirectional feature encoding table;
  • the blockchain on-chain module is used to upload the fruit pattern data obtained by the fruit pattern generation module after hash processing and signature, to obtain the block number of the blockchain where the fruit pattern is located, and to store the traceability information. After hashing the signature, the certificate is uploaded to the chain, and the block number of the traceability information is obtained.
  • the device for generating and uploading the fruit pattern map further includes a local application server, and the local application server is used to store the fruit pattern map, traceability information and corresponding block numbers of the fruit.
  • the image acquisition device includes a flexible clamping unit, a synchronous rotation unit, and an image acquisition unit, wherein,
  • Described flexible clamping unit has bidirectional telescopic function, is used for clamping both sides of fruit product according to the size, shape of different fruit product;
  • the synchronous rotation unit can realize 180-degree rotation, and is used for rotating the clamped fruit, so as to obtain two images of the pedicle and the hilum of each fruit;
  • the image acquisition unit is used for triggering image acquisition at certain intervals, and is used for respectively capturing an image of the stalk part and the hilum part of each fruit to be encoded.
  • the device for generating and winding the fruit pattern map further includes a tracing tool, and the tracing tool has a camera module, an arithmetic module and a communication module.
  • the traceability tool is used to take photos of the stalk or the hilum, obtain the fruit pattern through image processing, verify the authenticity of the product through comparison, and realize product traceability.
  • the purpose of the present invention is also to provide a system for reliable traceability of fruit products based on fruit pattern map and blockchain, including that the device includes a fruit pattern map generation module, a blockchain on-chain module, a blockchain server, and a local application.
  • the server and the client wherein, the fruit pattern map generation module obtains the image of the fruit surface (fruit stalk or fruit hilum), and performs edge positioning, ring extraction, normalization, image enhancement, feature extraction, feature encoding, and encoding table.
  • Merge and other processes generate a fruit pattern map;
  • the blockchain on-chain module hashes and signs the generated fruit pattern map through the pre-chain processing and on-chain processing, and then uploads the certificate to the chain to obtain the block of the blockchain where the fruit pattern map is located.
  • the traceability information is hashed and signed on the chain to store the certificate, and the block number of the traceability information is obtained;
  • the blockchain server is used to store the hash value of the fruit pattern map and the traceability information;
  • the client is used for shooting
  • the photo of the stalk or the navel is obtained by image processing to obtain the fruit pattern, hash the fruit pattern, and obtain the hash value stored on the blockchain through the block number for consistency comparison to verify the product. authenticity, and to achieve product traceability.
  • the fruit pattern map generation module and the blockchain uploading module in the above system are optionally the same as the corresponding components of the above-mentioned fruit pattern map generation and on-chain device.
  • the client can be a smart phone or other smart device with a camera module, a computing module and a communication module.
  • the blockchain server can use a public chain or a consortium chain.
  • the two-dimensional Gabor function is adopted, and its characteristic parameters are frequency and direction, which improves the accuracy of texture extraction and the efficiency of encoding.
  • the invention uses the appearance texture characteristics of the fruit itself to establish a feature map, solves the uniqueness and convenience of fruit identification, and realizes the first (identification) credibility of traceability; by sending the fruit texture map information to the blockchain, Further solve the problem of information tampering, and realize the second (information) credibility of traceability; by storing the fruit pattern information and business information on the blockchain and locally, and linking through the hash value, it is achieved.
  • trusted traceability saves blockchain storage space.
  • FIG. 1 is a schematic diagram of the working principle of an encoding and identification device based on fruit pattern map information according to a specific embodiment of the present invention.
  • FIG. 2 is a schematic diagram of the principle of an encoding and identification method based on fruit pattern map information according to a specific embodiment of the present invention.
  • FIG. 3 is a schematic flowchart of an encoding and identification method based on fruit pattern map information according to a specific embodiment of the present invention.
  • Fig. 4 is a schematic flow chart of part of the coding and identification method based on the fruit pattern map information according to the specific embodiment of the present invention.
  • FIG. 5 is a schematic diagram of the principle of a system for trusted traceability of fruit products based on a fruit pattern map and a blockchain according to a specific embodiment of the present invention.
  • Embodiment 1 The coding of fruit pattern map information
  • the specific embodiment of the present invention includes a kind of coding method based on fruit pattern map information, and the method comprises the following steps:
  • the fruit pattern map may further include a feature coding table of the pedicle part and/or the feature coding table of the hilum part.
  • FIG. 4 is a schematic flowchart of a part of the coding and identification method based on fruit pattern map information according to a specific embodiment of the present invention.
  • the steps of normalizing the image and converting it into a rectangular image of the pedicel part and the rectangular image of the hilum part include:
  • I(x, y) represents the ring image
  • the pedicle and the hilum After obtaining the images of the pedicle and the hilum, according to the difference between the grayscale of the pedicle and the hilum and the grayscale of the peripheral epidermis, locate the outer edges of the pedicle and the core of the hilum, and extract the pedicle and the hilum.
  • the central location of the umbilicus eg, the outer edges of the pedicle and the umbilical core are located using thresholding, mathematical morphological operations, and Hough transform. Because the outer edges of the pedicel and the umbilical core are generally roughly circular in shape, the center of the pedicled and the umbilical core can be determined by simple mathematical calculations.
  • the predetermined distances R1 and R2 are selected according to needs. First, the distance R1 is determined. When R1 is small, the fruit lines tend to be aggregated and not separated, and when R1 is large, the workload of calculation and coding may be too large. After R1 is determined, R2 can be selected to be a certain multiple of R1, for example, 1.2 to 1.8 times R1. Then extract the ring between the two concentric circles as the area to be processed;
  • the normalization process is actually a transformation mapping relationship from circular coordinates to rectangular coordinates. After transformation and mapping, it can be found that, for example, the radial texture of watermelon basically forms a striped texture in the rectangular coordinate system. .
  • histogram equalization is further adopted to enhance the image, so as to obtain a clearer texture; becomes:
  • N is the total number of pixels in the image to be enhanced
  • N(r k ) is the number of pixels whose gray level is rk in the image
  • k is the number of gray levels
  • T(r k ) is the number of pixels for the gray level rk Conversion function
  • S(r k ) is the converted gray level.
  • the present invention firstly enhances the rectangular image, and the reason for using histogram equalization is that in an image, the gray levels are often concentrated, such as dark
  • the component of the image histogram is concentrated at the lower end of the grayscale, while the component of the bright image histogram is biased to the end of the higher grayscale.
  • the uneven distribution on the grayscale will lead to a low degree of discrimination. If uniform coding is used, the coding accuracy will be improved. Lower, for this histogram equalization can make the gray distribution of the image more uniform.
  • the grayscale histogram of an image covers almost the entire grayscale value range, and except for the prominent number of individual grayscale values, the entire grayscale value distribution is approximately uniform, then the image has a larger grayscale value.
  • grayscale dynamic range and higher contrast while the details of the image are richer.
  • a transformation function can be obtained, and the input image can be used to achieve the above-mentioned effects. Therefore, the above-mentioned histogram equalization is adopted in the specific embodiment of the present invention as a pre-operation before image feature extraction. .
  • the described extracting features of the rectangular image of the pedicel part and the rectangular image of the fruit hilum part respectively, carry out image feature encoding, and the steps of obtaining the fruit pedicle part feature encoding table and the fruit hilum part feature encoding table include:
  • x 1 x cos ⁇ +y sin ⁇
  • y 1 -x sin ⁇ +y cos ⁇
  • the wavelength ⁇ is specified in pixels, for example, the wavelength value is usually greater than or equal to 2, but cannot be greater than one-fifth of the input image size, and ⁇ is the direction;
  • h Re , h Im represent the real and imaginary parts of the filtered feature parameters, respectively, fft represents the Fourier transform, and ifft represents the inverse Fourier transform.
  • a two-dimensional Gabor filter is used to extract image texture feature parameters. After inspection, it is found that the two-dimensional Gabor filter formed by the Gabor function has the characteristics of achieving optimal localization in the spatial and frequency domains at the same time, so it can well describe the selectivity corresponding to spatial frequency (scale), spatial position and direction. local structure information.
  • the frequency and direction representation of the two-dimensional Gabor filter is close to the representation of the human visual system for frequency and direction, so it can be used for texture representation and description.
  • a 2D Gabor filter is a product of a sine plane wave and a Gaussian kernel function.
  • the Gabor transform represents an optimization method for time-frequency analysis, while in the two-dimensional case it is a method for spatial-frequency domain analysis.
  • the window function determines its locality in the spatial domain, so the spatial domain information at different positions can be obtained by moving the center of the window. Therefore, a two-dimensional Gabor filter is used in the present invention.
  • the feature information of the two-dimensional space can be generated, and the obtained feature encoding table is relative to the one-dimensional image signal. It has a large amount of information and a high degree of discrimination, which is in line with the characteristics of the natural texture of the natural growth of the fruit.
  • T is the combined bidirectional feature coding table
  • T A and TB are the feature coding tables of the pedicel and hilum, respectively
  • X(i,j) is the value corresponding to the i-th row and the j-th column in the bidirectional coding table .
  • Embodiment 2 Recognition method based on fruit pattern atlas
  • Example 2 and 3 are schematic diagrams of the working principle and flow of this embodiment.
  • the method of Example 1 is used to obtain the fruit pattern atlas of each fruit, and the information of the fruit pattern is stored on the server.
  • the client takes pictures of the stem and hilum of the fruit to be identified, performs image feature encoding with the same method as above, matches the obtained feature encoding table with the stored feature encoding table, and is used for the identification of the feature information of the fruit pattern, including steps :
  • N is the number of bits of feature encoding
  • XOR represents XOR operation
  • P j and Q j respectively represent the jth bit of texture feature encoding P and Q;
  • the distance HD is greater than the predetermined first threshold, it belongs to different fruits, and if it is determined that the distance HD is smaller than the predetermined first threshold, it belongs to the same fruit.
  • the first threshold may be selected according to experience, or may be performed by means of machine learning. For example, encoding and identifying a certain number of fruits, adjusting the value of the first threshold, and checking the accuracy under different thresholds, thereby selecting the optimal solution of the first threshold.
  • the aforementioned encoding steps are first performed for the fruit to be identified to obtain the combined bidirectional feature encoding table, the feature encoding table of the pedicle part and the feature encoding table of the hilum part, and then the combined bidirectional feature encoding table is obtained.
  • the table, as well as the pedicle part feature coding table and the fruit hilum part feature coding table are matched and compared with the respective bidirectional feature coding tables, the pedicel part feature coding table and the fruit hilum part feature coding table stored in the storage, so as to carry out the identification of the fruit. operate.
  • the described fruit to be identified carries out image feature encoding, and the obtained feature encoding table is matched with the stored feature encoding table, and the identification for the feature information of the fruit pattern includes the steps:
  • the merged two-way feature encoding table identify the feature information of the fruit pattern according to the merged two-way feature encoding table. If it is judged that it belongs to the same fruit, then end the identification; otherwise, if it is determined that the difference between the distance and the first threshold is less than the predetermined difference, then further according to the fruit pedicle.
  • the partial feature encoding table and the fruit umbilical part feature encoding table are used to identify the feature information of the fruit pattern. If one of the determined distances obtained according to the fruit pedicle part feature encoding table or the fruit navel part feature encoding table is smaller than the predetermined first threshold, only It needs to be determined that the other of the distances is smaller than the second predetermined threshold, that is, it is judged that it belongs to the same fruit, and the identification is ended.
  • an obvious advantage of the present invention is that it has both accuracy and fault tolerance.
  • the combined bidirectional feature encoding table is used to identify the feature information of the fruit pattern, the fruit can be accurately identified, but when the fruit has a certain occurrence Therefore, although the distance HD determined according to the bidirectional feature encoding table is greater than the first predetermined threshold, when it is close to the first predetermined threshold, for example, when the difference between the determined distance HD and the first threshold is smaller than the predetermined difference, that is, , the distance HD determined according to the bidirectional feature coding table is increased by a limited amount from the predetermined first threshold.
  • the predetermined distance is set, it is only necessary to determine that the other of the distances is smaller than the second predetermined threshold to determine that it belongs to the same fruit, and the identification ends.
  • the second predetermined threshold value is greater than the first predetermined threshold value, indicating that as long as only one of the fruit pedicle and the fruit hilum is identified, the identification requirement of the other can be moderately released.
  • the combined bidirectional feature coding table can be used for matching first. If the combined bidirectional feature coding table can be matched, it means that the fruit to be identified is the same as the one stored in the database. The fruit is completely matched and can be used for fast matching; but if the local skin of the fruit is scratched, for example, one of the rectangular image of the fruit pedicle and the rectangular image of the hilum has changed in shape or pattern, resulting in a merged bidirectional feature.
  • the coding table cannot be matched, it can also be considered to use the feature coding table of the stem part and the feature coding table of the hilum part respectively for identification.
  • the coding and identification method of the present invention also has certain error tolerance, and can be compatible with detection efficiency and robustness.
  • FIG. 1 is a schematic diagram of the working principle of an encoding and identification device based on fruit pattern map information according to a specific embodiment of the present invention.
  • the specific embodiment of the present invention also includes an encoding and identification device based on the information of the pattern of the fruit pattern, which is based on the
  • the coding and identification device of the fruit pattern atlas information includes an image acquisition device, a fruit pattern atlas feature code generation device 4 and a fruit pattern atlas feature information identification device, wherein,
  • the image acquisition device is used to hold and rotate the fruit, and capture an image of the pedicle and the hilum of each fruit to be encoded;
  • the said atlas feature code generation device 4 is used for respectively graying out the image of the pedicel part and the part of the hilum, and normalizing the images to convert them into a rectangular image of the pedicel part and the rectangular image of the hilum part; and It is used to extract the features of the rectangular image of the pedicel part and the rectangular image of the hilum part, and encode the image features to obtain the feature coding table of the pedicel part and the umbilical part of the fruit.
  • the tables are combined and operated to obtain the combined bidirectional feature coding table; the bidirectional feature coding table, the pedicle part feature coding table and the fruit hilum part feature coding table are stored;
  • the fruit pattern feature information identification device is used for image feature encoding of the fruit to be recognized, and the obtained feature encoding table is matched with the stored feature encoding table for identification of the fruit pattern feature information.
  • the image acquisition device includes a flexible clamping unit 1, a synchronous rotation unit 2, and an image acquisition unit 3, wherein,
  • the flexible clamping unit has a bidirectional telescopic function, and is used to clamp both sides of the fruit according to the size and shape of different fruit;
  • the synchronous rotation unit can realize 180-degree rotation, and is used to rotate the clamped fruit, so as to obtain two images of the melon and the umbilicus of each fruit;
  • the image acquisition unit is used for triggering image acquisition at certain intervals, and is used for respectively acquiring an image of the stalk part and the hilum part of each fruit to be encoded.
  • the fruit pattern feature information identification device is used for image feature encoding of the fruit to be recognized, and the obtained feature encoding table is matched with the stored feature encoding table, which is used for the fruit pattern feature encoding.
  • Identification of information includes:
  • Embodiment 4 Fruit credible traceability method based on fruit pattern map and blockchain
  • the present embodiment provides a credible traceability method for fruit products based on a fruit texture atlas and a blockchain, which comprises the following steps:
  • E. Blockchain on-chain The fruit pattern map is hashed and signed, and then uploaded to the chain to store the certificate, and the block number of the blockchain where the fruit pattern map is located is obtained;
  • the step E includes:
  • Pre-processing on the chain perform a hashing process on the fruit pattern map to obtain a series of hash values bound to the fruit pattern map, use an asymmetric encryption algorithm to sign the hash value, and confirm with the public key and signature information
  • the sender holds the corresponding private key, thereby converting the fruit pattern into a way readable by the blockchain, and at the same time binding the sender's identity to the sent information through a signature;
  • On-chain processing The processed data is sent to the blockchain node to form a blockchain transaction and enter the on-chain stage; after receiving the transaction, each node in the blockchain first broadcasts the received transaction to other The nodes form a unified transaction pool; the transaction data contains the hash of the packaged transactions in the block, and the transactions need to be sorted according to a unified order; after confirming the block header and hash, calculate the block hash, and pass the previous block.
  • the hash and its own hash are connected to form a chain to complete the chaining process.
  • the method for credible traceability of fruit products based on the fruit pattern map and the blockchain also includes the following steps of hashing the traceability information of the fruit products, and then storing the hash values on the chain one by one to obtain the corresponding block number, and adding the corresponding block number.
  • the block number and the traceability information are updated to the local application server in a one-to-one correspondence.
  • the traceability information includes fruit variety information, planting information, cold chain logistics information, and the like.
  • the method of uploading traceability information is the same as above.
  • the described fruit credible traceability method based on fruit pattern map and block chain also comprises the steps:
  • Retrospective query use the above-mentioned method to obtain the fruit pattern spectrum of the fruit to be inspected, and use the described identification method to match the corresponding fruit pattern spectrum on the local application server through the fruit pattern spectrum of the fruit to be inspected, and obtain the area corresponding to the fruit.
  • Block number obtain the hash value of the fruit texture map stored on the blockchain through the block number, perform hash processing on the fruit texture map of the fruit stored on the local application server to obtain the local fruit texture map hash value, Consistently compare the hash value of the fruit pattern on the blockchain with the hash value of the local fruit pattern.
  • the traceability information of the fruit, the block number of the fruit pattern and the block number of the traceability information are obtained on the local application server through the fruit pattern map, and the fruit pattern map and the traceability information of the fruit to be inspected are hashed respectively, and are combined with each other.
  • Consumers can use a device with a camera function, such as a smart phone or other smart devices, to trace the source by taking images of the stem portion and/or the hilum portion of the fruit using the above-mentioned traceability method.
  • a device with a camera function such as a smart phone or other smart devices
  • the present invention provides a device for generating and uploading a fruit pattern map for the above method, the device comprising a fruit pattern map generation module and a blockchain uploading module, wherein,
  • the fruit pattern atlas generation module includes an image acquisition device and a fruit pattern atlas feature code generation device, and the image acquisition device is used to capture images of the stem portion and the hilum portion of each fruit to be encoded; the atlas feature encoding
  • the generating device is used to grayscale the image of the pedicel part and the part of the hilum, and normalize the images to convert them into a rectangular image of the pedicel and the rectangular image of the hilum; and is used to extract the pedicel respectively Part of the rectangular image and part of the fruit hilum part of the rectangular image features, image feature coding, to obtain the pedicle part of the feature coding table and the fruit hilum part of the feature coding table, for the fruit pedicle part of the feature coding table and the fruit hilum part of the feature coding table, perform a parallel operation to obtain The combined bidirectional feature encoding table;
  • the blockchain on-chain module is used to upload the fruit pattern data obtained by the fruit pattern generation module after hash processing and signature, to obtain the block number of the blockchain where the fruit pattern is located, and to store the traceability information. After processing the signature, upload the certificate to the chain to obtain the block number of the traceability information.
  • the device for generating and uploading the fruit pattern map further includes an application server, and the application server is used to store the fruit pattern map, traceability information and corresponding block numbers of the fruit.
  • the image acquisition device includes a flexible clamping unit, a synchronous rotation unit, and an image acquisition unit, wherein,
  • the flexible clamping unit has a bidirectional telescopic function, and is used to clamp both sides of the fruit according to the size and shape of different fruit;
  • the synchronous rotation unit can realize 180-degree rotation, and is used to rotate the clamped fruit, so as to obtain two images of the melon and the umbilicus of each fruit;
  • the image acquisition unit is used for triggering image acquisition at certain intervals, and is used for respectively capturing an image of the stalk part and the hilum part of each fruit to be encoded.
  • the device for generating and winding the fruit pattern map further includes a tracing tool, and the tracing tool has a camera module, an arithmetic module and a communication module.
  • the traceability tool is used to take photos of the stalk or the hilum, obtain the fruit pattern through image processing, verify the authenticity of the product through comparison, and realize product traceability.
  • Example 6 Fruit trusted traceability system based on fruit pattern and blockchain
  • this embodiment provides a system for credible traceability of fruit products based on a fruit pattern map and a blockchain, including that the device includes a fruit pattern map generation module 5 , a blockchain uploading module 6 , and a block chain.
  • the client 9 is used to take photos of the stem and the hilum, and obtain the traceability information of the fruit, the block number of the fruit pattern and the block number of the traceability information on the local application server through the fruit pattern map.
  • the fruit pattern map and traceability information of the fruit are hashed separately, and the hash value stored on the blockchain is obtained through the block number for consistency comparison to verify the authenticity of the product and achieve product traceability.
  • the fruit pattern map generation module and the blockchain uploading module in the system of this embodiment are optionally the same as the corresponding components of the fruit pattern map generation and on-chain device in Embodiment 5.
  • the client can be a smart phone or other smart device with a camera module, a computing module and a communication module.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Security & Cryptography (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Power Engineering (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)

Abstract

本发明公开了一种基于果纹图谱和区块链的果品可信追溯方法及装置。本发明通过获取单个果品的果蒂部分和果脐部分的图像，对图像进行灰度化、归一化处理，转换为矩形图像；分别提取矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于这两个特征编码表进行并操作，获得合并后的双向特征编码表，形成具有唯一性的果纹图谱，通过将果纹图谱进行处理，将果纹图谱信息及相关信息上链存证，用户通过智能终端采用同样的算法获得待检果品的果纹图谱，经处理后与链上信息进行核验达到可信追溯的目的。本发明实现了果类识别的唯一性和便捷性，通过链上存证，解决信息被篡改的问题，达到了可信追溯的目的。

Description

一种基于果纹图谱和区块链的果品可信追溯方法及装置 技术领域
本发明涉及植物特征提取、识别以及区块链技术，特别是涉及到基于水果天然纹路的信息编码、识别和区块链追溯技术。
背景技术
追溯作为质量管理的有效措施从20世纪80年代被引入食品工业至今，欧盟、美国、加拿大、澳大利亚等国相继建立了农产品及食品追溯系统。近年来，随着我国对食品质量安全的日益重视及民众安全消费意识的不断提升，根据供应链的特点及质量安全监管的现实需求，建立“从田间到餐桌”的全过程质量安全追溯体系，已成为确保民众“舌尖上安全”的迫切需要。
对于果蔬、肉类、水产品等产品，相关专利从标识方法、信息采集、数据传递等角度出发，形成了较为成熟的技术体系。如：
发明专利-用于蔬菜追溯的编码方法及所用编码系统(申请号/专利号：2011101649608)，公开了一种用于蔬菜追溯的编码系统；还公开了相应的一种用于蔬菜追溯的编码方法。
发明专利-动物个体溯源的方法及其设备(申请号/专利号：200910199977.X)，通过给分配有现行码号的动物个体固定一个能储存动物标识识别码编号的电子标识，并将现行码号转换成动物标识识别码编号，将转换成的动物标识识别码编号储存入所述电子标识中，读出所述电子标识中的国际动物代码编号，用于访问国际和国内动物信息数据。
发明专利-水产养殖产品质量安全全程管理与追溯方法及系统(申请号/专利号：200610113644.7)，提供了一种水产养殖产品质量安全管理与追溯方法及系统，它通过固化无公害水产养殖标准生产数据库，采用危害分析与关键控制点HACCP体系，通过对水产养殖产品(鱼、虾、蟹)与从育苗到放养， 从饲料投喂到药物使用养殖生产全过程管理，提供了与全球统一编码(EAN/UCC)相接轨的以条码自动识别技术为支撑的水产养殖管理与追溯解决方案。
这些方法采用了条码以及RFID等作为识别标签，但是对于水果等果品，如果附加条码信息以进行编码、识别和溯源，则会存在各种困难和缺点，首先，果品自身形状不规则，贴制条码或RFID标签易脱落，而且果品不规则表面不利于条码的读取；其次，果品表面贴制标签容易污染果品表皮，影响果品品质，也容易产生质量安全问题；另外，单个果品贴制标签，既增加了标签等耗材的成本，也增加了人力成本；另外，标签是外在附着物，对于高端果品来说，可能存在标签人为调换的风险，相同标签信息不一定意味着相同的水果；此外，追溯相关信息还存在着造假的可能。
发明内容
为了解决上述问题，本发明第一个目的在于，提出果纹图谱信息的编码方法，用于生成果纹图谱，其能够利用果品自身的果纹特征，构建基于果纹特征的图谱的编码方法，建立每个果品的“身份证”；
本发明的第二个目的在于提供一种基于果纹图谱的识别方法，通过客户端对待检果品果纹图谱与存储在服务器端的果纹图谱进行比对，进而实现果品的单果识别；
本发明的第三个目的在于提供一种基于果纹图谱和区块链的果品可信追溯方法，利用区块链技术将果品果纹谱图存储到区块链上，实现果纹图谱数据的不可篡改；
本发明的第四个目的在于提供用于上述方法的果纹图谱生成与上链装置；
本发明的目的还在于提供一种基于果纹图谱和区块链的果品可信追溯的系统。
为了实现上述目标，本发明采取了如下的技术方案：
一种果纹图谱信息的编码方法，其特征在于，该方法包括步骤：
A、获取待编码果品的果蒂部分和果脐部分的图像；
B、分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；
C、分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表，形成果纹图谱。该果纹图谱还可以进一步包括果蒂部分特征编码表和/或果脐部分特征编码表。
本发明的特点之一在于利用水果的天然纹路作为基准，将该天然纹路中提取出图像特征，并进行图像特征编码，一般说来，没有纹路完全相同的两个水果，因此只要控制果蒂部分和果脐部分图像具有足够的分辨率，可以将世上所有的水果都编出不同的特征编码出来，因此本发明的技术效果在于准确性、唯一性。
因为水果的天然纹路不易篡改，无法复制，即使水果生长发生了形状改变也基本保持原有纹路信息，因此本发明具有能够防伪、便于溯源的突出优点。对于纹路比较显著的水果，例如西瓜、哈密瓜、苹果等水果的编码和识别，特别适合。
另外，本发明的特点之一在于，分别利用了水果果蒂部分矩形图像和果脐部分矩形图像的特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表，因此进行特征编码表匹配以识别水果时，可以首先利用合并后的双向特征编码表进行匹配，如果合并后的双向特征编码表能匹配，则说明待识别水果与数据库中所存储的水果完全匹配，能够用于快速匹配；但如果发生了水果的局部外皮擦伤，例如水果果蒂部分矩 形图像和果脐部分矩形图像中的一个发生了形状或花纹改变，导致合并后的双向特征编码表不能匹配时，也可以考虑分别利用果蒂部分特征编码表和果脐部分特征编码表来进行识别，当仅有一种特征编码表能够匹配并符合其他条件时，也能被视作识别成功，因此本发明的编码和识别方法还具有一定的容错性，能够兼容检测效率和鲁棒性。
另外，所述对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像的步骤包括：
B1、在果蒂部分图像和果脐部分图像中，根据果蒂和果脐的灰度与外围表皮的灰度存在的差异，对果蒂和果脐核心的外边缘进行定位，并提取果蒂和果脐的中心位置；
B2、分别以提取的果蒂和果脐的中心位置为圆心，以中心到边缘的预定距离R1和R2为半径绘制同心圆，提取两个同心圆之间的圆环为待处理区域；
B3、对待处理区域进行归一化处理为：
I(x(r,θ),y(r,θ))→I(r,θ)，
其中，
Figure PCTCN2021103363-appb-000001
其中，I(x,y)表示圆环图像；(r,θ)表示归一化后的极坐标，r∈[0,1],θ∈[0°,360°]，当r＝0时，表示I(x(r,θ),y(r,θ))为圆环图像内边缘的像素点；当r＝1时，表示I(x(r,θ),y(r,θ))为圆环图像外边缘的像素点；对于待处理区域中的每一个点(x i,y i)，分别考虑与中心点(x 0,y 0)的关系，确定其(r,θ)，并以r和θ为直角坐标，将圆环图像变换为直角坐标下的矩形图像I(r,θ)。
另外，对于转换后的果蒂部分矩形图像和果脐部分矩形图像，分别进一步采取直方图均衡化来增强图像，用于获得更清晰的纹理；所述采取直方图均衡化为：
Figure PCTCN2021103363-appb-000002
其中N是所述待增强图像像素的总数，N(r k)为图像出现灰度级为r k的像素数，k为灰度级数，T(r k)为对于灰度级r k的转换函数，S(r k)为转换后的灰度级。
另外，所述分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表的步骤包括：
C1、首先用哈尔小波变换提取各通道的能量均值μ及方差σ，使用K均值聚类方法进行聚类，得到小样本集的圆环图像；
C2、使用二维Gabor滤波提取圆环图像的纹理信息，得到相应的纹理特征参数；其中所述二维Gabor滤波的表达式为：
Figure PCTCN2021103363-appb-000003
其中，x 1＝x cosθ+y sinθ,y 1＝-x sinθ+y cosθ；波长λ以像素为单位指定，λ表示Gabor核函数中余弦函数的波长参数，以像素为单位指定，通常大于等于2，但不能大于输入图像尺寸的五分之一；而θ为方向；指定了Gabor函数并行条纹的方向，它的取值为0到360度。另外，σ表示Gabor函数的高斯因子的标准差。
C3、得到纹理特征参数后，判断其系数的实部和虚部的正负进行量化编码，具体为：
Figure PCTCN2021103363-appb-000004
Figure PCTCN2021103363-appb-000005
其中，h Re,h Im分别表示为滤波后特征参数的实部和虚部，fft表示傅里叶变换，ifft表示为傅里叶反变换。其中，G(f)表示滤波器，与前述的G(x,y)对应；I(r)表示归一化后的图像。
另外，所述对于果蒂部分特征编码表和果脐部分特征编码表进行并操作， 获得合并后的双向特征编码表为：
C4、T＝T A∪T B＝{X (i,j)|X (i,j)∈T A或X (i,j)∈T B}，
其中T为合并后的双向特征编码表，T A、T B分别为果蒂部分和果脐部分的特征编码表，X(i,j)为双向编码表中第i行第j列对应的值。
为实现本发明的第二个目的，进一步本发明提供一种基于果纹图谱的识别方法，该方法采用上述方法获得果纹图谱，以及步骤D、对待识别的果品进行图像特征编码，将获得的特征编码表与所述果纹图谱的特征编码表进行匹配，用于果纹图谱特征信息的识别。
另外，所述对待识别的水果进行图像特征编码，将获得的特征编码表与存储的特征编码表进行匹配，用于果纹图谱特征信息的识别包括步骤：
D1、使用基于汉明距离的分类器进行匹配，其距离计算公式为：
Figure PCTCN2021103363-appb-000006
其中，N为特征编码位数，XOR表示异或运算，P j、Q j分辨表示纹理特征编码P、Q的第j位；
确定距离大于预定第一阈值，则属于不同的果品，确定距离小于预定第一阈值，则属于同一果品。
另外，所述对待识别的水果进行图像特征编码，将获得的特征编码表与存储的特征编码表进行匹配，用于果纹图谱特征信息的识别包括步骤：
首先根据合并后的双向特征编码表进行果纹图谱特征信息的识别，若判断属于同一果品，则结束识别；否则如果确定距离与第一阈值之差小于预定差值时，则从双向特征编码表中分别提取果蒂部分特征编码表和果脐部分特征编码表，并进一步分别根据果蒂部分特征编码表和果脐部分特征编码表进行果纹图谱特征信息的识别，如果依据果蒂部分特征编码表或果脐部分特征编码表得到的确定距离中的一个小于预定第一阈值，则只需确定距离中的另一个小于第二预定阈值即判断属于同一果品，结束识别。
为实现本发明的第三个目的，本发明提供一种基于果纹图谱和区块链的 果品可信追溯方法，其包括采用上述的方法获得果品的果纹图谱，以及步骤：
E、区块链上链：将所述果纹图谱经哈希处理签名后上链存证，得到该果纹图谱所在区块链的区块号；
F、将所述区块号与该区块号果纹图谱信息存储到本地应用服务器。
具体地，所述步骤E包括：
E1、上链前处理：将果纹图谱进行一次哈希处理，得到与果纹图谱绑定的一串哈希值，采用非对称加密算法对哈希值进行签名，通过公钥和签名信息确认发送者持有对应的私钥，从而将果纹图谱转换成区块链可读的方式，同时通过签名将发送者的身份与发送信息绑定；
E2、上链处理：处理完后的数据发送到区块链节点，形成一笔区块链交易进入上链阶段；在收到交易后，区块链各节点将接收到的交易先广播到其他节点，形成一个统一的交易池；交易数据包含该区块里打包交易的哈希，交易需要根据统一的顺序排序；在确认区块头和哈希之后，计算区块哈希，通过前一区块哈希和自身哈希相连形成链条，完成上链过程。
进一步地所述基于果纹图谱和区块链的果品可信追溯方法，还包括将该果品的追溯信息分别经哈希处理后将哈希值逐条上链存证获得相应区块号，并将该区块号及追溯信息一一对应更新到本地应用服务器。所述追溯信息包括果品品种信息、种植信息、冷链物流信息等等。追溯信息的上链方法同上。
进一步地所述基于果纹图谱和区块链的果品可信追溯方法，还包括步骤：
G、追溯查询：采用上述的方法获得待检果品的果纹图谱，采用所述的识别方法通过待检果品的果纹图谱在本地应用服务器上匹配相应的果纹图谱并获取该果品对应的区块号，通过区块号获取存储在区块链上的果纹图谱哈希值，将本地应用服务器上存储的该果品的果纹图谱进行哈希处理获得本地果纹图谱哈希值，将所述区块链上的果纹图谱哈希值与所述本地果纹图谱哈希值进行一致性比对。
具体地，所述追溯查询采用上述方法获得待检果品的果纹图谱，通过果纹图谱在本地应用服务器上获得该果品的追溯信息、果纹图谱区块号以及追溯信息区块号，对待检果品的果纹图谱以及追溯信息分别进行哈希计算，并与通过区块号获取存储在区块链上的哈希值进行一致性对比，判断追溯果品是否为与该果纹图谱对应的果品，并判断追溯信息是否被篡改。。
消费者可以利用带有拍照功能的装置例如智能手机或其他智能设备，通过拍摄果品的果蒂部分和/或果脐部分的图像用上述追溯方法进行溯源。
为实现上述第四个目的，本发明提供一种用于上述方法的果纹图谱生成与上链装置，所述装置包括果纹图谱生成模块、区块链上链模块，其中，
所述果纹图谱生成模块包括图像获取装置、果纹图谱特征编码生成装置，所述图像获取装置用于对每个待编码果品的果蒂部分和果脐部分分别摄取图像；所述图谱特征编码生成装置用于分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；以及用于分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表；
所述区块链上链模块用于将果纹图谱生成模块获得的果纹图谱数据经哈希处理签名后上链存证，得到果纹图谱所在区块链的区块号，以及将追溯信息经哈希处理签名后上链存证，得到追溯信息的区块号。
进一步，所述果纹图谱生成与上链装置还包括本地应用服务器，所述本地应用服务器用于存储果品的果纹图谱、追溯信息以及相应区块号。
进一步，所述图像获取装置包括柔性夹持单元、同步转动单元、图像获取单元，其中，
所述柔性夹持单元具有双向伸缩功能，用于根据不同果品的大小、形状 夹持果品的两侧；
所述同步转动单元能实现180度旋转，用于将夹持住的果品进行转动，以便于获取每个果品的果蒂和果脐部分两幅图像；
所述图像获取单元用于每隔一定间隔触发图像获取，用于对每个待编码果品的果蒂部分和果脐部分分别摄取一幅图像。
进一步地，所述果纹图谱生成与上链装置还包括追溯工具，所述追溯工具具有摄像模块、运算模块和通讯模块。追溯工具用于拍摄果蒂或果脐处照片，通过图像处理获得其果纹图谱，通过比对验证产品的真伪性，并实现产品追溯。
本发明的目的还在于提供一种基于果纹图谱和区块链的果品可信追溯的系统，包括所述装置包括果纹图谱生成模块、区块链上链模块、区块链服务器、本地应用服务器以及客户端，其中，果纹图谱生成模块通过获取果品表面(果蒂或果脐处)图像，通过边缘定位、圆环提取、归一化处理、图像增强、特征提取、特征编码、编码表合并等流程生成果纹图谱；区块链上链模块通过上链前处理和上链处理将生成的果纹图谱哈希处理签名后上链存证，得到果纹图谱所在区块链的区块号，以及将追溯信息经哈希处理签名后上链存证，得到追溯信息的区块号；区块链服务器用于存储果纹图谱哈希值以及追溯信息哈希值；客户端用于拍摄果蒂或果脐处照片，通过图像处理获得其果纹图谱，对果纹图谱进行哈希计算，并于通过区块号获取存储在区块链上的哈希值进行一致性对比，验证产品的真伪性，并实现产品追溯。
上述系统中的果纹图谱生成模块、区块链上链模块可选地与上述果纹图谱生成与上链装置相应部件相同。所述客户端可以是具有摄像模块、运算模块和通讯模块的智能手机或者其他智能设备。
在本发明中，区块链服务器可以采用公有链或者联盟链。
本发明的技术效果包括如下：
1.利用果类自身具有的外表纹理特性建立图谱特征信息，解决果类识别的唯一性和便捷性。所述果纹信息不易遗失，难以篡改，能充分保障水果识别和溯源过程中的准确性。
2.通过果蒂部分特征编码表和果脐部分特征编码表建立合并后的双向特征编码表，提高识别的准确性，首先根据合并后的双向特征编码表进行果纹图谱特征信息的识别，若判断属于同一果品，则结束识别；否则如果确定距离与第一阈值之差小于预定差值时，则进一步分别根据果蒂部分特征编码表和果脐部分特征编码表进行果纹图谱特征信息的识别，综合考虑识别结果进一步判断是否属于同一果品；因此首先提高了准确度，同时保证了一定的纠错能力。
3.采用二维Gabor函数，其特征参数有频率和方向，提升了纹理提取的精度和编码的效率。
本发明利用果类自身具有的外表纹理特性建立特征图谱，解决果类识别的唯一性和便捷性，实现追溯的第一重(标识)可信；通过将果纹图谱信息发送到区块链，进一步解决信息被篡改的问题，实现追溯的第二重(信息)可信；通过将果纹图谱信息和业务信息分别在区块链上存储和本地存储，并通过哈希值链接，即达到了可信追溯的目的，又节省了区块链存储空间。
附图说明
图1为根据本发明具体实施方式中基于果纹图谱信息的编码和识别装置工作原理的示意图。
图2为根据本发明具体实施方式中基于果纹图谱信息的编码和识别方法的原理示意图。
图3为根据本发明具体实施方式中基于果纹图谱信息的编码和识别方法的流程示意图。
图4为根据本发明具体实施方式中基于果纹图谱信息的编码和识别方法 的部分流程示意图。
图5为根据本发明具体实施方式中一种基于果纹图谱和区块链的果品可信追溯的系统的原理示意图。
具体实施方式
下面结合附图，对本发明作详细说明。
以下公开详细的示范实施例。然而，此处公开的具体结构和功能细节仅仅是出于描述示范实施例的目的。
然而，应该理解，本发明不局限于公开的具体示范实施例，而是覆盖落入本公开范围内的所有修改、等同物和替换物。在对全部附图的描述中，相同的附图标记表示相同的元件。
参阅附图，本说明书所附图式所绘示的结构、比例、大小等，均仅用以配合说明书所揭示的内容，以供熟悉此技术的人士了解与阅读，并非用以限定本发明可实施的限定条件，故不具技术上的实质意义，任何结构的修饰、比例关系的改变或大小的调整，在不影响本发明所能产生的功效及所能达成的目的下，均应仍落在本发明所揭示的技术内容得能涵盖的范围内。同时，本说明书中所引用的位置限定用语，亦仅为便于叙述的明了，而非用以限定本发明可实施的范围，其相对关系的改变或调整，在无实质变更技术内容下，当亦视为本发明可实施的范畴。
同时应该理解，如在此所用的术语“和/或”包括一个或多个相关的列出项的任意和所有组合。另外应该理解，当部件或单元被称为“连接”或“耦接”到另一部件或单元时，它可以直接连接或耦接到其他部件或单元，或者也可以存在中间部件或单元。此外，用来描述部件或单元之间关系的其他词语应该按照相同的方式理解(例如，“之间”对“直接之间”、“相邻”对“直接相邻”等)。
实施例1 果纹图谱信息的编码
本发明具体实施方式中包括一种基于果纹图谱信息的编码方法，该方法包括以下步骤：
A、夹持水果并进行转动，对每个待编码水果的果蒂部分和果脐部分分别摄取一幅图像；
B、分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；
C、分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表，形成果纹图谱。该果纹图谱进一步还可包括果蒂部分特征编码表和/或果脐部分特征编码表。
另外，图4为根据本发明具体实施方式中基于果纹图谱信息的编码和识别方法的部分流程示意图。如图所示，在本发明具体实施方式中，所述对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像的步骤包括：
B1、在果蒂部分图像和果脐部分图像中，根据果蒂和果脐的灰度与外围表皮的灰度存在的差异，对果蒂和果脐核心的外边缘进行定位，并提取果蒂和果脐的中心位置；
B2、分别以提取的果蒂和果脐的中心位置为圆心，以中心到边缘的预定距离R1和R2为半径绘制同心圆，提取两个同心圆之间的圆环为待处理区域；
B3、对待处理区域进行归一化处理为：
I(x(r,θ),y(r,θ))→I(r,θ)，
其中，
Figure PCTCN2021103363-appb-000007
其中，I(x,y)表示圆环图像；(r,θ)表示归一化后的极坐标，r∈[0,1], θ∈[0°,360°]，当r＝0时，表示I(x(r,θ),y(r,θ))为圆环图像内边缘的像素点；当r＝1时，表示I(x(r,θ),y(r,θ))为圆环图像外边缘的像素点；对于待处理区域中的每一个点(x i,y i)，分别考虑与中心点(x 0,y 0)的关系，确定其(r,θ)，并以r和θ为直角坐标，将圆环图像变换为直角坐标下的矩形图像I(r,θ)。
获得果蒂部分图像和果脐部分图像后，根据果蒂和果脐的灰度与外围表皮的灰度存在的差异，对果蒂和果脐核心的外边缘进行定位，并提取果蒂和果脐的中心位置，例如采用阈值法、数学形态学运算以及Hough变换对果蒂和果脐核心的外边缘进行定位。因为果蒂和果脐核心的外边缘一般形状大致为圆形，因此通过简单的数学计算即可确定果蒂和果脐核心的圆心。
所述预定距离R1和R2根据需要选取，首先确定距离R1，当R1较小时，水果纹路容易出现聚集未分开，而当R1较大时，可能导致计算、编码的工作量过大。确定R1后，可以选择R2为R1的一定倍数，例如1.2～1.8倍R1。进而提取两个同心圆之间的圆环为待处理区域；
所述归一化处理，实际上是一种圆形坐标向直角坐标下的变换映射关系，通过变换映射后可以发现，例如西瓜的放射状纹理基本上形成了直角坐标系中的条带状的纹理。
另外，本发明具体实施方式中，对于转换后的果蒂部分矩形图像和果脐部分矩形图像，分别进一步采取直方图均衡化来增强图像，用于获得更清晰的纹理；所述采取直方图均衡化为：
Figure PCTCN2021103363-appb-000008
其中N是所述待增强图像像素的总数，N(r k)为图像出现灰度级为r k的像素数，k为灰度级数，T(r k)为对于灰度级r k的转换函数，S(r k)为转换后的灰度级。
为了便于对果蒂部分矩形图像和果脐部分矩形图像进行图形特征提取，本发明首先对于矩形图像进行了增强，采用直方图均衡化的原因在于一副图 像中，往往灰度比较集中，例如暗图像直方图的分量集中在灰度较低的一端，而亮图像直方图分量偏向于灰度较高的一端，在灰度上的分布不均匀会导致区分度低，如果使用均匀编码则编码精度较低，为此直方图均衡化可以让图像的灰度分布更加均匀。如果一幅图像的灰度直方图几乎覆盖了整个灰度的取值范围，并且除了个别灰度值的个数较为突出，整个灰度值分布近似于均匀分布，那么这幅图像就具有较大的灰度动态范围和较高的对比度，同时图像的细节更为丰富。仅仅依靠输入图像的直方图信息，就可以得到一个变换函数，利用该变换函数可以将输入图像达到上述效果，因此本发明具体实施方式中采用了上述直方图均衡化作为图像特征提取前的预操作。
另外，本发明具体实施方式中，所述分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表的步骤包括：
C1、首先用哈尔小波变换提取各通道的能量均值μ及方差σ，使用K均值聚类方法进行聚类，得到小样本集的圆环图像；
C2、使用二维Gabor滤波提取圆环图像的纹理信息，得到相应的纹理特征参数；其中所述二维Gabor滤波的表达式为：
Figure PCTCN2021103363-appb-000009
其中，x 1＝x cosθ+y sinθ,y 1＝-x sinθ+y cosθ；波长λ以像素为单位指定，例如波长取值通常大于等于2，但不能大于输入图像尺寸的五分之一，而θ为方向；
C3、得到纹理特征参数后，判断其系数的实部和虚部的正负进行量化编码，具体为：
Figure PCTCN2021103363-appb-000010
Figure PCTCN2021103363-appb-000011
其中，h Re,h Im分别表示为滤波后特征参数的实部和虚部，fft表示傅里叶变换，ifft表示为傅里叶反变换。
本发明具体实施方式中，特别使用了二维Gabor滤波器来提取图像纹理特征参数。经过检验发现，用Gabor函数形成的二维Gabor滤波器具有在空间域和频率域同时取得最优局部化的特性，因此能够很好地描述对应于空间频率(尺度)、空间位置及方向选择性的局部结构信息。二维Gabor滤波器的频率和方向表示接近人类视觉系统对于频率和方向的表示，因此可以被用于纹理表示和描述。实际上在空域，一个2维的Gabor滤波器是一个正弦平面波和高斯核函数的乘积。
在一维情况中，Gabor变换代表着时频分析的优化方法，而二维情况中则是空间频域分析的方法。对于图像来说，窗函数决定了它在空域的局部性，所以可以通过移动窗口的中心來获得不同位置的空间域信息。因此本发明中使用了二维Gabor滤波器，相对于某些情况下的一维Gabor滤波器而言，能够生成二维空间的特征信息，获得的特征编码表相对于一维图像信号而言，信息量大、区分度高，符合水果自然生长的天然纹理的特点。
另外，本发明具体实施方式中，所述对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表为：
C4、T＝T A∪T B＝{X (i,j)|X (i,j)∈T A或X (i,j)∈T B}，
其中T为合并后的双向特征编码表，T A、T B分别为果蒂部分和果脐部分的特征编码表，X(i,j)为双向编码表中第i行第j列对应的值。
实施例2 基于果纹图谱的识别方法
图2和3是本实施例的工作原理及流程示意图。本例采用实施例1的方法获得各个果品的果纹图谱，并将果纹图谱信息存储于服务器上。客户端拍摄待识别果品的果蒂和果脐图像，用上述相同的方法进行图像特征编码，将获得的特征编码表与存储的特征编码表进行匹配，用于果纹图谱特征信息的 识别包括步骤：
D1、使用基于汉明距离的分类器进行匹配，其距离计算公式为：
Figure PCTCN2021103363-appb-000012
其中，N为特征编码位数，XOR表示异或运算，P j、Q j分辨表示纹理特征编码P、Q的第j位；
确定距离HD大于预定第一阈值，则属于不同的果品，确定距离小于预定第一阈值，则属于同一果品。
所述第一阈值可以根据经验选取，也可以通过机器学习的方式来进行。例如通过一定数量的水果进行编码和识别，调整所述第一阈值的取值，并检查不同阈值下的准确率，由此选择第一阈值的最优解。
本发明具体实施方式中，对于要识别的水果首先执行前述编码步骤，得到合并后的双向特征编码表，以及果蒂部分特征编码表和果脐部分特征编码表，然后将合并后的双向特征编码表，以及果蒂部分特征编码表和果脐部分特征编码表与存储在存储的各个双向特征编码表、果蒂部分特征编码表和果脐部分特征编码表进行匹配比较，由此进行水果的识别操作。
另外，本发明具体实施方式中，所述对待识别的水果进行图像特征编码，将获得的特征编码表与存储的特征编码表进行匹配，用于果纹图谱特征信息的识别包括步骤：
首先根据合并后的双向特征编码表进行果纹图谱特征信息的识别，若判断属于同一果品，则结束识别；否则如果确定距离与第一阈值之差小于预定差值时，则进一步分别根据果蒂部分特征编码表和果脐部分特征编码表进行果纹图谱特征信息的识别，如果依据果蒂部分特征编码表或果脐部分特征编码表得到的确定距离中的一个小于预定第一阈值，则只需确定距离中的另一个小于第二预定阈值即判断属于同一果品，结束识别。
如前所述，本发明一个明显的优点在于准确性与容错性都具备，当使用合并后的双向特征编码表进行果纹图谱特征信息的识别，能够准确地识别水 果，但是当水果发生了一定的损伤，导致根据双向特征编码表确定的距离HD虽然大于第一预定阈值，但接近所述第一预定阈值时，例如所述确定距离HD与第一阈值之差小于预定差值时，亦即，根据双向特征编码表确定的距离HD比预定第一阈值增大的量较为有限。则进一步分别根据果蒂部分特征编码表和果脐部分特征编码表进行果纹图谱特征信息的识别，如果依据果蒂部分特征编码表或果脐部分特征编码表得到的确定距离中的一个小于预定第一阈值，则只需确定距离中的另一个小于第二预定阈值即判断属于同一果品，结束识别。所述第二预定阈值大于第一预定阈值，表明只要果蒂和果脐中只要一个被识别，则可以适度放开另一个的识别要求。
因此本发明的方法进行特征编码表匹配以识别水果时，可以首先利用合并后的双向特征编码表进行匹配，如果合并后的双向特征编码表能匹配，则说明待识别水果与数据库中所存储的水果完全匹配，能够用于快速匹配；但如果发生了水果的局部外皮擦伤，例如水果果蒂部分矩形图像和果脐部分矩形图像中的一个发生了形状或花纹改变，导致合并后的双向特征编码表不能匹配时，也可以考虑分别利用果蒂部分特征编码表和果脐部分特征编码表来进行识别，当仅有一种特征编码表能够匹配并符合其他条件时，也能被视作识别成功，因此本发明的编码和识别方法还具有一定的容错性，能够兼容检测效率和鲁棒性。
实施例3 基于果纹图谱信息的编码和识别装置
图1为根据本发明具体实施方式中基于果纹图谱信息的编码和识别装置工作原理的示意图。如图所示，与本发明具体实施方式中的基于果纹图谱信息的编码和识别方法相适应，本发明具体实施方式中还包括一种基于果纹图谱信息的编码和识别装置，所述基于果纹图谱信息的编码和识别装置包括图像获取装置、果纹图谱特征编码生成装置4和果纹图谱特征信息识别装置，其中，
所述图像获取装置用于夹持水果并进行转动，对每个待编码水果的果蒂部分和果脐部分分别摄取一幅图像；
所述图谱特征编码生成装置4用于分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；以及用于分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表；存储双向特征编码表、果蒂部分特征编码表和果脐部分特征编码表；
所述果纹图谱特征信息识别装置用于对待识别的水果进行图像特征编码，将获得的特征编码表与存储的特征编码表进行匹配，用于果纹图谱特征信息的识别。
另外，本发明具体实施方式中，所述图像获取装置包括柔性夹持单元1、同步转动单元2、图像获取单元3，其中，
所述柔性夹持单元具有双向伸缩功能，用于根据不同果品的大小、形状夹持果品的两侧；
所述同步转动单元能实现180度旋转，用于将夹持住的果品进行转动，以便于获取每个果品的瓜蒂和瓜脐部分两幅图像；
所述图像获取单元用于每隔一定间隔触发图像获取，用于对每个待编码水果的果蒂部分和果脐部分分别摄取一幅图像。
另外，本发明具体实施方式中，所述果纹图谱特征信息识别装置用于对待识别的水果进行图像特征编码，将获得的特征编码表与存储的特征编码表进行匹配，用于果纹图谱特征信息的识别包括：
首先根据合并后的双向特征编码表进行果纹图谱特征信息的识别，若判断属于同一果品，则结束识别；否则根据预定条件满足时，分别根据果蒂部 分特征编码表和果脐部分特征编码表进行果纹图谱特征信息的识别，综合考虑识别结果进一步判断是否属于同一果品。
因此，图谱特征信息的识别，综合考虑识别结果进一步判断是否属于同一果品。
实施例4 基于果纹图谱和区块链的果品可信追溯方法
本实施例提供一种基于果纹图谱和区块链的果品可信追溯方法，其包括采用实施例1的方法获得果品的果纹图谱，以及步骤：
E、区块链上链：将所述果纹图谱经哈希处理签名后上链存证，得到该果纹图谱所在区块链的区块号；
F、将所述区块号与该区块号果纹图谱信息存储到本地应用服务器。
具体地，所述步骤E包括：
E1、上链前处理：将果纹图谱进行一次哈希处理，得到与果纹图谱绑定的一串哈希值，采用非对称加密算法对哈希值进行签名，通过公钥和签名信息确认发送者持有对应的私钥，从而将果纹图谱转换成区块链可读的方式，同时通过签名将发送者的身份与发送信息绑定；
E2、上链处理：处理完后的数据发送到区块链节点，形成一笔区块链交易进入上链阶段；在收到交易后，区块链各节点将接收到的交易先广播到其他节点，形成一个统一的交易池；交易数据包含该区块里打包交易的哈希，交易需要根据统一的顺序排序；在确认区块头和哈希之后，计算区块哈希，通过前一区块哈希和自身哈希相连形成链条，完成上链过程。
进一步地所述基于果纹图谱和区块链的果品可信追溯方法，还包括将该果品的追溯信息分别经哈希处理后将哈希值逐条上链存证获得相应区块号，并将该区块号及追溯信息一一对应更新到本地应用服务器。所述追溯信息包括果品品种信息、种植信息、冷链物流信息等等。追溯信息的上链方法同上。
进一步地，所述基于果纹图谱和区块链的果品可信追溯方法，还包括步 骤：
G、追溯查询：采用上述的方法获得待检果品的果纹图谱，采用所述的识别方法通过待检果品的果纹图谱在本地应用服务器上匹配相应的果纹图谱并获取该果品对应的区块号，通过区块号获取存储在区块链上的果纹图谱哈希值，将本地应用服务器上存储的该果品的果纹图谱进行哈希处理获得本地果纹图谱哈希值，将所述区块链上的果纹图谱哈希值与所述本地果纹图谱哈希值进行一致性比对。
进一步地，通过果纹图谱在本地应用服务器上获得该果品的追溯信息、果纹图谱区块号以及追溯信息区块号，对待检果品的果纹图谱以及追溯信息分别进行哈希计算，并与通过区块号获取存储在区块链上的哈希值进行一致性对比，判断追溯果品是否为与该果纹图谱对应的果品，并判断追溯信息是否被篡改。
消费者可以利用带有拍照功能的装置例如智能手机或其他智能设备，通过拍摄果品的果蒂部分和/或果脐部分的图像用上述追溯方法进行溯源。
实施例5 果纹图谱生成与上链装置
本发明提供一种用于上述方法的果纹图谱生成与上链装置，所述装置包括果纹图谱生成模块、区块链上链模块，其中，
所述果纹图谱生成模块包括图像获取装置、果纹图谱特征编码生成装置，所述图像获取装置用于对每个待编码果品的果蒂部分和果脐部分分别摄取图像；所述图谱特征编码生成装置用于分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；以及用于分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表；
所述区块链上链模块用于将果纹图谱生成模块获得的果纹图谱数据经哈希处理签名后上链存证，得到果纹图谱所在区块链的区块号，以及将追溯信息经处理签名后上链存证，得到追溯信息的区块号。
进一步，所述果纹图谱生成与上链装置还包括应用服务器，所述应用服务器用于存储果品的果纹图谱、追溯信息以及相应区块号。
进一步，所述图像获取装置包括柔性夹持单元、同步转动单元、图像获取单元，其中，
所述柔性夹持单元具有双向伸缩功能，用于根据不同果品的大小、形状夹持果品的两侧；
所述同步转动单元能实现180度旋转，用于将夹持住的果品进行转动，以便于获取每个果品的瓜蒂和瓜脐部分两幅图像；
所述图像获取单元用于每隔一定间隔触发图像获取，用于对每个待编码果品的果蒂部分和果脐部分分别摄取一幅图像。
进一步地，所述果纹图谱生成与上链装置还包括追溯工具，所述追溯工具具有摄像模块、运算模块和通讯模块。追溯工具用于拍摄果蒂或果脐处照片，通过图像处理获得其果纹图谱，通过比对验证产品的真伪性，并实现产品追溯。
实施例6 基于果纹图谱和区块链的果品可信追溯系统
如图5所示，本实施例提供一种基于果纹图谱和区块链的果品可信追溯的系统，包括所述装置包括果纹图谱生成模块5、区块链上链模块6、区块链服务器7、本地应用服务器8以及客户端9，其中，果纹图谱生成模块5通过获取果品表面(果蒂或果脐处)图像，通过边缘定位、圆环提取、归一化处理、图像增强、特征提取、特征编码、编码表合并等流程生成果纹图谱；区块链上链模块6通过上链前处理和上链处理将生成的果纹图谱哈希处理签名后上链存证，得到果纹图谱所在区块链的区块号，以及将追溯信息经哈希处 理签名后上链存证，得到追溯信息的区块号；区块链服务器用于存储果纹图谱哈希值以及追溯信息哈希值；客户端9用于拍摄果蒂及果脐处照片，通过果纹图谱在本地应用服务器上获得该果品的追溯信息、果纹图谱区块号以及追溯信息区块号，对待检果品的果纹图谱以及追溯信息分别进行哈希计算，并于通过区块号获取存储在区块链上的哈希值进行一致性对比，验证产品的真伪性，并实现产品追溯。
本实施例系统中的果纹图谱生成模块、区块链上链模块可选地与实施例5中的果纹图谱生成与上链装置相应部件相同。所述客户端可以是具有摄像模块、运算模块和通讯模块的智能手机或者其他智能设备。
上述说明示出并描述了本发明的若干优选实施例，但如前所述，应当理解本发明并非局限于本说明书所披露的形式，不应看作是对其他实施例的排除，而可用于各种其他组合、修改和环境，并能够在本说明书所述发明构想范围内，通过上述教导或相关领域的技术或知识进行改动。而本领域人员所进行的改动和变化不脱离本发明的精神和范围，则都应在本发明所附权利要求的保护范围内。

Claims (18)

  1. 一种果纹图谱信息的编码方法，其特征在于，该方法包括步骤：
    A、获取待编码水果的果蒂部分和果脐部分的图像；
    B、分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；
    C、分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表，形成果纹图谱。
  2. 根据权利要求1中所述的方法，其特征在于，所述对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像的步骤包括：
    B1、在果蒂部分图像和果脐部分图像中，根据果蒂和果脐的灰度与外围表皮的灰度存在的差异，对果蒂和果脐核心的外边缘进行定位，并提取果蒂和果脐的中心位置；
    B2、分别以提取的果蒂和果脐的中心位置为圆心，以中心到边缘的预定距离R1和R2为半径绘制同心圆，提取两个同心圆之间的圆环为待处理区域；
    B3、对待处理区域进行归一化处理为：
    I(x(r,θ),y(r,θ))→I(r,θ)，
    其中，
    Figure PCTCN2021103363-appb-100001
    其中，I(x,y)表示圆环图像；(r,θ)表示归一化后的极坐标，r∈[0,1],θ∈[0°,360°]，当r＝0时，表示I(x(r,θ),y(r,θ))为圆环图像内边缘的像素点；当r＝1时，表示I(x(r,θ),y(r,θ))为圆环图像外边缘的像素点；对于待处理区域中的每一个点(x i,y i)，分别考虑与中心点(x 0,y 0)的关系，确定其(r,θ)，并 以r和θ为直角坐标，将圆环图像变换为直角坐标下的矩形图像I(r,θ)。
  3. 根据权利要求2中所述的方法，其特征在于，对于转换后的果蒂部分矩形图像和果脐部分矩形图像，分别进一步采取直方图均衡化来增强图像，用于获得更清晰的纹理；所述采取直方图均衡化为：
    Figure PCTCN2021103363-appb-100002
    其中N是所述待增强图像像素的总数，N(r k)为图像出现灰度级为r k的像素数，k为灰度级数，T(r k)为对于灰度级r k的转换函数，S(r k)为转换后的灰度级。
  4. 根据权利要求1中所述的方法，其特征在于，所述分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表的步骤包括：
    C1、首先用哈尔小波变换提取各通道的能量均值μ及方差σ，使用K均值聚类方法进行聚类，得到小样本集的圆环图像；
    C2、使用二维Gabor滤波提取圆环图像的纹理信息，得到相应的纹理特征参数；其中所述二维Gabor滤波的表达式为：
    Figure PCTCN2021103363-appb-100003
    其中，x 1＝x cosθ+y sinθ,y 1＝-x sinθ+y cosθ；波长λ以像素为单位指定，θ为方向；
    C3、得到纹理特征参数后，判断其系数的实部和虚部的正负进行量化编码，具体为：
    Figure PCTCN2021103363-appb-100004
    Figure PCTCN2021103363-appb-100005
    其中，h Re,h Im分别表示为滤波后特征参数的实部和虚部，fft表示傅里叶变换，ifft表示为傅里叶反变换。
  5. 根据权利要求1中所述的方法，其特征在于，所述对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表为：
    C4、T＝T A∪T B＝{X (i,j)|X (i,j)∈T A或X (i,j)∈T B}，
    其中T为合并后的双向特征编码表，T A、T B分别为果蒂部分和果脐部分的特征编码表，X(i,j)为双向编码表中第i行第j列对应的值。
  6. 一种基于果纹图谱的识别方法，其包括采用权利要求1～5所述的方法获得果纹图谱，以及，步骤
    D、对待识别的果品进行图像特征编码，将获得的特征编码表与所述果纹图谱的特征编码表进行匹配，用于果纹图谱特征信息的识别。
  7. 根据权利要求6所述的方法，其特征在于，所述步骤D包括：
    D1、使用基于汉明距离的分类器进行匹配，其距离计算公式为：
    Figure PCTCN2021103363-appb-100006
    其中，N为特征编码位数，XOR表示异或运算，P j、Q j分辨表示纹理特征编码P、Q的第j位；
    确定距离大于预定第一阈值，则属于不同的果品，确定距离小于预定第一阈值，则属于同一果品。
  8. 根据权利要求7中所述的方法，其特征在于，所述步骤D包括：
    首先根据合并后的双向特征编码表进行果纹图谱特征信息的识别，若判断属于同一果品，则结束识别；否则如果确定距离与预定第一阈值之差小于预定差值时，则从双向特征编码表中分别提取果蒂部分特征编码表和果脐部分特征编码表，并进一步分别根据果蒂部分特征编码表和果脐部分特征编码表进行果纹图谱特征信息的识别，如果依据果蒂部分特征编码表或果脐部分特征编码表得到的确定距离中的一个小于预定第一阈值，则只需确定距离中的另一个小于预定第二阈值即判断属于同一果品，结束识别。
  9. 一种基于果纹图谱和区块链的果品双重可信追溯方法，其包括采用权利要求1～5所述的方法获得果品的果纹图谱，以及步骤：
    E、区块链上链：将所述果纹图谱经哈希处理后将哈希值上链存证，得到该果纹图谱所在区块链的区块号；
    F、将所述区块号与该区块号果纹图谱信息存储到本地应用服务器。
  10. 根据权利要求9所述的方法，其特征在于，所述步骤E包括：
    E1、上链前处理：将果纹图谱进行一次哈希处理，得到与果纹图谱绑定的一串哈希值，采用非对称加密算法对哈希值进行签名，通过公钥和签名信息确认发送者持有对应的私钥，从而将果纹图谱转换成区块链可读的方式，同时通过签名将发送者的身份与发送信息绑定；
    E2、上链处理：处理完后的数据发送到区块链节点，形成一笔区块链交易进入上链阶段；在收到交易后，区块链各节点将接收到的交易先广播到其他节点，形成一个统一的交易池；交易数据包含该区块里打包交易的哈希，交易需要根据统一的顺序排序；在确认区块头和哈希之后，计算区块哈希，通过前一区块哈希和自身哈希相连形成链条，完成上链过程。
  11. 根据权利要求9或10所述的方法，其特征在于，还包括将该果品的追溯信息分别经哈希处理后将哈希值逐条上链存证获得相应区块号，并将该区块号及追溯信息一一对应更新到本地应用服务器。
  12. 根据权利要求9或10所述的方法，其特征在于，还包括步骤：
    G、追溯查询：采用权利要求1～5任一项所述的方法获得待检果品的果纹图谱，采用权利要求6～8所述的方法通过待检果品的果纹图谱在本地应用服务器上匹配相应的果纹图谱并获取该果品对应的区块号，通过区块号获取存储在区块链上的果纹图谱哈希值，将本地应用服务器上存储的该果品的果纹图谱进行哈希处理获得本地果纹图谱哈希值，将所述区块链上的果纹图谱哈希值与所述本地果纹图谱哈希值进行一致性比对。
  13. 根据权利要求12所述的方法，其特征在于，还包括步骤：
    G1、追溯查询：采用权利要求1～6任一项所述的方法获得待检果品的果 纹图谱，通过所述果纹图谱在本地应用服务器上获得该果品的追溯信息、果纹图谱区块号以及追溯信息区块号，对待检果品的果纹图谱以及追溯信息分别进行哈希计算，并与通过区块号获取存储在区块链上的哈希值进行一致性对比，判断追溯果品是否为与该果纹图谱对应的果品，并判断追溯信息是否被篡改。
  14. 一种用于权利要求9～13任一项所述方法的果纹图谱生成与上链装置，其特征在于，所述装置包括果纹图谱生成模块、区块链上链模块，其中，
    所述果纹图谱生成模块包括图像获取装置、果纹图谱特征编码生成装置，所述图像获取装置用于对每个待编码果品的果蒂部分和果脐部分分别摄取图像；所述图谱特征编码生成装置用于分别对果蒂部分图像和果脐部分图像进行灰度化，并对图像进行归一化处理，转换为果蒂部分矩形图像和果脐部分矩形图像；以及用于分别提取果蒂部分矩形图像和果脐部分矩形图像特征，进行图像特征编码，获得果蒂部分特征编码表和果脐部分特征编码表，对于果蒂部分特征编码表和果脐部分特征编码表进行并操作，获得合并后的双向特征编码表；
    所述区块链上链模块用于将果纹图谱生成模块获得的果纹图谱数据经哈希处理签名后上链存证，得到果纹图谱所在区块链的区块号。
  15. 根据权利要求14所述的果纹图谱生成与上链装置，其特征在于，所述区块链上链模块还包括将追溯信息经哈希处理签名后上链存证，得到追溯信息的区块号
  16. 根据权利要求15所述的装置，其特征在于，还包括本地应用服务器，所述本地应用服务器用于存储果品的果纹图谱、追溯信息以及相应区块号。
  17. 根据权利要求15所述的装置，其特征在于，所述图像获取装置包括柔性夹持单元、同步转动单元、图像获取单元，其中，
    所述柔性夹持单元具有双向伸缩功能，用于根据不同果品的大小、形状 夹持果品的两侧；
    所述同步转动单元能实现180度旋转，用于将夹持住的果品进行转动，以便于获取每个果品的果蒂和果脐部分两幅图像；
    所述图像获取单元用于每隔一定间隔触发图像获取，用于对每个待编码果品的果蒂部分和果脐部分分别摄取一幅图像。
  18. 根据权利要求15所述的装置，其特征在于，还包括追溯工具，所述追溯工具具有摄像模块、运算模块和通讯模块。
PCT/CN2021/103363 2020-12-09 2021-06-30 一种基于果纹图谱和区块链的果品可信追溯方法及装置 WO2022121290A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/698,142 US20220207789A1 (en) 2020-12-09 2022-03-18 Credible fruit traceability method and device based on fruit texture atlas and blockchain

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011447866.9A CN112488233B (zh) 2020-12-09 2020-12-09 一种基于果纹图谱信息的编码和识别方法及装置
CN202011447866.9 2020-12-09

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/698,142 Continuation US20220207789A1 (en) 2020-12-09 2022-03-18 Credible fruit traceability method and device based on fruit texture atlas and blockchain

Publications (1)

Publication Number Publication Date
WO2022121290A1 true WO2022121290A1 (zh) 2022-06-16

Family

ID=74941542

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/103363 WO2022121290A1 (zh) 2020-12-09 2021-06-30 一种基于果纹图谱和区块链的果品可信追溯方法及装置

Country Status (3)

Country Link
US (1) US20220207789A1 (zh)
CN (1) CN112488233B (zh)
WO (1) WO2022121290A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115983874A (zh) * 2023-02-17 2023-04-18 江苏秀圆果信息科技有限公司 酒类防伪溯源方法及其系统
CN116976919A (zh) * 2023-09-25 2023-10-31 国品优选(北京)品牌管理有限公司 基于区块链的口服液防伪溯源方法及系统
CN117347312A (zh) * 2023-12-06 2024-01-05 华东交通大学 基于多光谱结构光的柑橘连续检测方法和设备

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112488233B (zh) * 2020-12-09 2021-12-17 中国农业科学院农业资源与农业区划研究所 一种基于果纹图谱信息的编码和识别方法及装置
CN116562716B (zh) * 2023-07-10 2023-09-19 北京佳格天地科技有限公司 一种基于物联网的产品质量溯源系统以及方法
CN116976917B (zh) * 2023-07-31 2024-05-24 金景(海南)科技发展有限公司 基于区块链技术的农业品牌存证体系建设方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4741042A (en) * 1986-12-16 1988-04-26 Cornell Research Foundation, Inc. Image processing system for detecting bruises on fruit
CN103679144A (zh) * 2013-12-05 2014-03-26 东南大学 一种基于计算机视觉的复杂环境下果蔬识别方法
CN105004737A (zh) * 2015-07-14 2015-10-28 浙江大学 基于自适应改进型梯度信息的水果表面缺陷检测方法
CN108389062A (zh) * 2018-05-10 2018-08-10 江南大学 一种基于图像处理的水果防伪溯源系统及方法
CN109493082A (zh) * 2018-09-25 2019-03-19 西安纸贵互联网科技有限公司 一种农产品区块链溯源方法及装置
CN112488233A (zh) * 2020-12-09 2021-03-12 中国农业科学院农业资源与农业区划研究所 一种基于果纹图谱信息的编码和识别方法及装置

Family Cites Families (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5845002A (en) * 1994-11-03 1998-12-01 Sunkist Growers, Inc. Method and apparatus for detecting surface features of translucent objects
US9886945B1 (en) * 2011-07-03 2018-02-06 Reality Analytics, Inc. System and method for taxonomically distinguishing sample data captured from biota sources
JP6316569B2 (ja) * 2013-11-01 2018-04-25 株式会社ブレイン 物品識別システムとそのプログラム
CN104636716B (zh) * 2014-12-08 2018-04-13 宁波工程学院 绿色果实识别方法
CN106529547B (zh) * 2016-10-14 2019-05-03 天津师范大学 一种基于完备局部特征的纹理识别方法
US10839503B2 (en) * 2017-01-26 2020-11-17 ClariFruit System and method for evaluating fruits and vegetables
CN107527362A (zh) * 2017-08-14 2017-12-29 西安交通大学 一种基于图像纹理特征指标的苹果口感定性鉴别方法
US11645835B2 (en) * 2017-08-30 2023-05-09 Board Of Regents, The University Of Texas System Hypercomplex deep learning methods, architectures, and apparatus for multimodal small, medium, and large-scale data representation, analysis, and applications
CN108305148A (zh) * 2018-02-06 2018-07-20 赵航 一种基于物联网的农产品贸易系统
CN108509927B (zh) * 2018-04-09 2021-09-07 中国民航大学 一种基于局部对称图结构的手指静脉图像识别方法
CN108596250B (zh) * 2018-04-24 2019-05-14 深圳大学 图像特征编码方法、终端设备及计算机可读存储介质
CN109784204B (zh) * 2018-12-25 2023-04-07 江苏大学 一种用于并联机器人的堆叠串类水果主果梗识别和提取方法
US10949974B2 (en) * 2019-02-28 2021-03-16 Iunu, Inc. Automated plant disease detection
CN111626085A (zh) * 2019-02-28 2020-09-04 中科院微电子研究所昆山分所 一种检测方法、装置、设备及介质
CN110163629A (zh) * 2019-04-09 2019-08-23 南京新立讯科技股份有限公司 一种商品溯源码生成及查询方法和装置
US20220252568A1 (en) * 2019-07-15 2022-08-11 Clarifruit Ltd. Means and methods for scoring vegetables and fruits
CN111351766A (zh) * 2020-02-27 2020-06-30 浙江大学 一种快速识别南瓜种子身份的方法
US20230143130A1 (en) * 2020-04-10 2023-05-11 Agropeeper Technologies Private Limited System and method for identifying fruit shelf life
CN111721728B (zh) * 2020-07-16 2023-02-21 陈皓 一种水果在线检测装置及其使用方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4741042A (en) * 1986-12-16 1988-04-26 Cornell Research Foundation, Inc. Image processing system for detecting bruises on fruit
CN103679144A (zh) * 2013-12-05 2014-03-26 东南大学 一种基于计算机视觉的复杂环境下果蔬识别方法
CN105004737A (zh) * 2015-07-14 2015-10-28 浙江大学 基于自适应改进型梯度信息的水果表面缺陷检测方法
CN108389062A (zh) * 2018-05-10 2018-08-10 江南大学 一种基于图像处理的水果防伪溯源系统及方法
CN109493082A (zh) * 2018-09-25 2019-03-19 西安纸贵互联网科技有限公司 一种农产品区块链溯源方法及装置
CN112488233A (zh) * 2020-12-09 2021-03-12 中国农业科学院农业资源与农业区划研究所 一种基于果纹图谱信息的编码和识别方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHEN, XUEXIN ET AL.: "Research on Fruit Recognition Algorithm Based on Multi-Color and Local Texture", JOURNAL OF QINGDAO UNIVERSITY(ENGINEERING & TECHNOLOGY EDITION), vol. 34, no. 3, 31 August 2019 (2019-08-31), pages 52 - 58, XP055942118 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115983874A (zh) * 2023-02-17 2023-04-18 江苏秀圆果信息科技有限公司 酒类防伪溯源方法及其系统
CN116976919A (zh) * 2023-09-25 2023-10-31 国品优选(北京)品牌管理有限公司 基于区块链的口服液防伪溯源方法及系统
CN116976919B (zh) * 2023-09-25 2024-01-02 国品优选(北京)品牌管理有限公司 基于区块链的口服液防伪溯源方法及系统
CN117347312A (zh) * 2023-12-06 2024-01-05 华东交通大学 基于多光谱结构光的柑橘连续检测方法和设备
CN117347312B (zh) * 2023-12-06 2024-04-26 华东交通大学 基于多光谱结构光的柑橘连续检测方法和设备

Also Published As

Publication number Publication date
CN112488233B (zh) 2021-12-17
CN112488233A (zh) 2021-03-12
US20220207789A1 (en) 2022-06-30

Similar Documents

Publication Publication Date Title
WO2022121290A1 (zh) 一种基于果纹图谱和区块链的果品可信追溯方法及装置
CN113283446B (zh) 图像中目标物识别方法、装置、电子设备及存储介质
Bereta et al. Local descriptors and similarity measures for frontal face recognition: a comparative analysis
Nguyen et al. Leaf based plant identification system for android using surf features in combination with bag of words model and supervised learning
CN112069891B (zh) 一种基于光照特征的深度伪造人脸鉴别方法
CN106096348B (zh) 一种基于多维码的证卡验证系统及方法
CN102542243A (zh) 一种基于lbp图像和分块编码的虹膜特征提取方法
Yan et al. Gabor surface feature for face recognition
Muhammad et al. Copy move image forgery detection method using steerable pyramid transform and texture descriptor
CN106022782A (zh) 一种虹膜支付系统
CN110647820A (zh) 基于特征空间超分辨映射的低分辨率人脸识别方法
Zhao et al. Facial expression recognition based on fusion of Gabor and LBP features
Zhao et al. Visible-infrared person re-identification based on frequency-domain simulated multispectral modality for dual-mode cameras
Tong et al. Local dominant directional symmetrical coding patterns for facial expression recognition
CN115169375B (zh) 基于ar与枪球联动的高位物料可视化方法
CN202694370U (zh) 一种基于数字图像处理的多人脸识别系统
Pavithra et al. Texture image classification and retrieval using multi-resolution radial gradient binary pattern
CN105512677B (zh) 基于Hash编码的遥感图像分类方法
Sivanarain et al. Ear recognition based on local texture descriptors
CN115170616B (zh) 人员轨迹分析方法、装置、终端及存储介质
Zhang et al. A scalable and efficient multi-label CNN-based license plate recognition on spark
Chen et al. Palmprint classification using contourlets
Yuan et al. Dual-encoded features from both spatial and Curvelet domains for image smoke recognition
CN106156787B (zh) 多模态湿地生态生境场景核空间溯源方法及装置
Shengli et al. Scene recognition of photovoltaic panels based on model migration and convolution neural network

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21902002

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21902002

Country of ref document: EP

Kind code of ref document: A1