CN113052234A - Jade classification method based on image features and deep learning technology - Google Patents
Jade classification method based on image features and deep learning technology Download PDFInfo
- Publication number
- CN113052234A CN113052234A CN202110316436.1A CN202110316436A CN113052234A CN 113052234 A CN113052234 A CN 113052234A CN 202110316436 A CN202110316436 A CN 202110316436A CN 113052234 A CN113052234 A CN 113052234A
- Authority
- CN
- China
- Prior art keywords
- jade
- image
- jades
- detection result
- detection
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 239000010977 jade Substances 0.000 title claims abstract description 127
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000013135 deep learning Methods 0.000 title claims abstract description 19
- 238000005516 engineering process Methods 0.000 title claims abstract description 17
- 238000001514 detection method Methods 0.000 claims abstract description 104
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 20
- 238000004458 analytical method Methods 0.000 claims abstract description 18
- 238000013136 deep learning model Methods 0.000 claims abstract description 9
- 239000000284 extract Substances 0.000 claims abstract description 3
- 238000012545 processing Methods 0.000 claims description 24
- 239000000463 material Substances 0.000 claims description 20
- 238000007781 pre-processing Methods 0.000 claims description 11
- -1 Duyu Substances 0.000 claims description 10
- 229910052640 jadeite Inorganic materials 0.000 claims description 10
- 239000013078 crystal Substances 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 238000007635 classification algorithm Methods 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 230000017105 transposition Effects 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 241000304460 Sinocrassula yunnanensis Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
- G06V10/30—Noise filtering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a jade classification method based on image characteristics and a deep learning technology, which comprises an image acquisition module, a jade detection module, a detection result analysis module and a result output module, wherein the image acquisition module acquires jade images through acquisition equipment, preprocesses the input images, extracts the characteristics of the images by using a deep learning model algorithm, detects and classifies jades according to the characteristics of the images and the jade, and determines the number of jades in the current image, the positions of the jades, the types of the jades and the confidence coefficient, thereby realizing the intelligent classification of the jades. The invention adopts a deep learning method to solve the outstanding problems of low identification accuracy, low identification efficiency, high classification cost and the like in the prior art. In order to make up for the defects of the prior art, the invention provides a simple and efficient jade classification method based on image features and a deep learning technology.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to a jade classification method based on image features and a deep learning technology.
Background
In recent years, with the rapid development of artificial intelligence technology, especially the falling of advanced science technologies such as artificial intelligence based on machine learning and deep learning in various industries, the mode of "AI + traditional industry" is leading a new innovation, which can only be completed manually in the past, and is gradually replaced by artificial intelligence technology at present, so that semi-automation, automation and intellectualization are gradually realized.
In the jade industry, jades are mainly classified into jadeites and nephrite, and different types of jades exist in jadeites and nephrite, wherein jadeites are represented by jades, and nephrite is represented by nephrite. Chinese jades are various in types, including jadeite, Hetian jade, yellow dragon jade, Russian material, seed material, Xiuyuyan jade, Duyu, crystal and the like. In the market transaction of the jade industry, the classification of the raw material types of the jade is mainly judged by manpower, and specific instruments and equipment are required to identify the jade which is difficult to identify. However, in the classification method in the prior art, judgment by naked eyes is not only inefficient and not high in accuracy, but classification by special instruments and equipment is high in cost, time-consuming and greatly limited in processing capacity, and is difficult to adapt to marketization requirements.
Disclosure of Invention
In view of the defects in the background art, the invention relates to a jade classification method based on image features and a deep learning technology. In order to make up for the defects of the prior art, the invention provides a simple and efficient jade classification method based on image features and a deep learning technology.
The invention relates to a jade classification method based on image characteristics and a deep learning technology, which comprises an image acquisition module, a jade detection module, a detection result analysis module and a result output module, wherein the image acquisition module acquires jade images through acquisition equipment, preprocesses the input images, extracts the characteristics of the images by using a deep learning model algorithm, detects and classifies jades according to the characteristics of the images and the jade, and determines the number of jades in the current image, the positions of the jades, the types of the jades and the confidence coefficient, thereby realizing the intelligent classification of the jades.
By adopting the scheme, the intelligent recognition and classification are carried out on the features of the images by using the pre-trained deep learning model algorithm, so that the recognition accuracy and efficiency are effectively improved.
Furthermore, the acquisition module acquires images by adopting acquisition equipment, wherein the images comprise visible light images and non-visible light images acquired by professional equipment; the jade detection module comprises image preprocessing, model detection, jade number detection and detection result reliability judgment; the detection result analysis module comprises jade quantity analysis, position analysis, category analysis and confidence coefficient analysis, and the result output module comprises detection result judgment completion and information uploading.
By adopting the scheme, the image to be detected can be obtained through various ways, including all images shot or analyzed by video streams.
Further, the acquisition module of the acquisition module includes, but is not limited to, a client camera, a professional camera, a scanner, a digital camera, a microscope, an infrared emitter, a flaw detector, and the like.
Further, the image preprocessing and image processing methods include, but are not limited to, scaling, translation, transposition, mirroring, rotation, normalization, dimensionality reduction, denoising, equalization, smoothing, and the like.
By adopting the scheme, the picture processing is more convenient, and the same management and identification speed after the picture processing is high.
Further, the method for training the model for model detection may use an object detection algorithm, an image segmentation algorithm or a classification algorithm.
By adopting the scheme, a large amount of picture materials of jades such as jade, Hetian jade, yellow dragon jade, Russian material, seed material, Xiuyu jade, Duyu, crystal and the like are collected, classified and calibrated according to different jade categories, model parameters are trained and learned by using an image detection or classification or segmentation algorithm through computer equipment, and the required pre-training model is finally obtained after training is finished.
Further, the method comprises the following implementation steps:
s1: acquiring an image to be detected of the jade through image acquisition equipment; s2: image preprocessing, namely processing the format of the image to achieve the specified size and format; s3: model detection, namely calculating the number and the class probability value of the jades contained in the image by adopting a pre-trained model; s4: judging whether a detection result exists or not according to whether the number of the detected jades is not 0 or not, and executing S7 if the detection result exists, namely the number of the detected jades is not less than 1; if no detection result exists, namely the number of the detected jades is 0, executing S5; s5: whether the threshold value is exceeded or not can be determined according to factors such as scene requirements, computing power performance of a software and hardware platform, fault tolerance rate of jade missing detection and the like, if the number of times of no detection result does not exceed the threshold value N0, S6 is executed, image resolution, brightness, saturation and the like are adjusted, and detection is continued; if the number of times of no detection result exceeds the threshold value N0, executing S1, wherein the input image does not contain jade, and continuing to process the next image; s6: image processing, namely performing operations such as denoising, normalization and the like on an image and processing the format of the image to achieve the specified size and format; s7: analyzing the detection result, namely analyzing the jades one by one, wherein the jade probability, the number of jades, the image positions of the jades, the width and the height and the jade type probability are included; s8: whether the confidence coefficient exceeds a threshold value or not, if the confidence coefficient of the detection result does not exceed the threshold value N1, executing S5, processing the image and then detecting again; if the confidence of the detection result exceeds the threshold value N1, the detection result is reliable, and S9 is executed; s9: whether the detection result is processed or not is finished, if the detection result is not processed, S7 is executed, and the next jade image is processed continuously; if the detection result is processed, S10 is executed, and the result is reported. S10: reporting results, wherein the reported results comprise the number, the position, the confidence coefficient and the jade probability information of the jade.
By adopting the scheme, the method has the advantages of high jade classification accuracy, high recognition efficiency, strong scene adaptability, good robustness, low cost and the like.
Further, the probability of jade in S7 and S10 is a probability value belonging to jadeite, nephrite, jadeite, yellow dragon jade, russian material, seed material, hsiuyen jade, Duyu, crystal, and other categories.
Further, the image formats, files, and resolutions in S1, S2, and S6 are not limited by upper and lower limits of the size, nor by the image format.
Further, the models in S3 include, but are not limited to, the Yolo series, MobileNet series, SSD, Pelee, RCNN, FastRCN, FasterRCNN, Mask R-CNN, VGGNet, ResNet, GoogleNet, FCN, SetNet, ReSeg, CRF, MRF, and the like, and variant algorithms thereof.
The invention has the following beneficial effects:
1. the innovation of the jade classification method is realized;
2. the novel jade can be adapted after retraining by adjusting the classification of the model;
3. the number of jades is not limited, and 1 or any more images of jades can be processed;
4. the data source of the collected image is not limited, for example, the image is shot by a mobile phone camera, a professional camera, a scanner, a digital camera or an image analyzed by a video stream, and the image is collected by professional equipment such as a microscope, an infrared emission instrument, a flaw detector, and the like. The image format is not limited, various pictures such as jpg, png, tif and gif, video frames, screenshots, scanning pieces and the like can be supported, and the size of the acquired image file is not limited by upper and lower limits.
Drawings
The invention is further illustrated with reference to the following figures and examples.
Fig. 1 is a schematic block diagram of an embodiment of the present invention.
Fig. 2 is a block flow diagram of an embodiment of the invention.
Reference numeral, 1, an image acquisition module; 2. a jade detection module; 3. a detection result analysis module; 4. and a result output module.
Detailed Description
While the embodiments of the present invention will be described and illustrated in detail with reference to the accompanying drawings, it is to be understood that the invention is not limited to the specific embodiments disclosed, but is intended to cover various modifications, equivalents, and alternatives falling within the scope of the invention as defined by the appended claims.
For the convenience of understanding the embodiments of the present invention, the following description will be further explained by taking specific embodiments as examples with reference to the drawings, and the embodiments are not to be construed as limiting the embodiments of the present invention.
The embodiment 1 of the invention refers to fig. 1 and fig. 2, and comprises an image acquisition module 1, a jade detection module 2, a detection result analysis module 3 and a result output module 4. The image acquisition module 1 acquires the picture through the picture acquisition equipment, and the image acquired by the acquisition equipment comprises images obtained by shooting or analyzing video streams of a mobile phone camera, a professional camera, a scanner, a digital camera and the like and images acquired by professional equipment such as a microscope, an infrared emission instrument, a flaw detector and the like. The image format is not limited, and includes but not limited to various pictures such as jpg, png, tif, gif, etc., video frames, screenshots, scannings, etc., and there is no upper and lower limit to the size of the acquired image file.
The jade detection module 2 is mainly used for preprocessing an input image and calling a deep learning model to perform jade detection, and the preprocessing is used for preparing for detection by using the deep learning model in the next step, so that the detection accuracy is improved. The jade detection module 2 comprises image preprocessing, model detection, detection result judgment, threshold exceeding judgment and image processing. The image preprocessing and the image processing are a processing mode of multiple images, and the mode includes but is not limited to geometric transformation such as translation, transposition, mirroring, rotation and scaling, standardization processing, normalization, dimension reduction, equalization, smoothing, denoising and other common methods in the field of image processing. The model detection adopts a pre-trained model, and the method for training the model can be a target detection algorithm, an image segmentation algorithm or a classification algorithm, such as SSD, Yolov5 and the like, and the specific method comprises the following steps: firstly, a large amount of picture materials of jades such as jades, Hetian jade, jadeite, yellow dragon jade, Russian material, seed material, Xiuyan jade, Duyu, crystal and the like are collected, classification and calibration are carried out according to different jade categories, model parameter training learning is carried out by using image detection or classification or segmentation algorithms through computer equipment, and a required pre-training model is finally obtained after training is finished. The output result of the model detection comprises the number of jades in the image, and the probability value of each jade belonging to the categories of jadeite, Hetian jade, yellow dragon jade, Russian material, seed material, Hsiuyen jade, Duyu, crystal, and the like. The judgment of the detection result is used for judging whether the judgment of the detection result is carried out according to whether the number of the detected jades is not 0. Exceeding the threshold value, judging whether the times of no detection result exceed a threshold value N0, if so, adjusting the resolution, brightness, saturation and the like of the image, and continuing to detect; if not, it is determined that the image does not contain the jade, and the next image is executed. Threshold N0 determines the mode: the method can be determined according to factors such as scene requirements, computing power performance of software and hardware platforms, fault tolerance rate of jade missing detection and the like, and for example, the threshold N0 is 3.
The detection result analysis module 3 analyzes the detected jades one by one, and the detection result of the deep learning model of each jade mainly comprises the following information:
firstly, detecting a jade area in an original image, and the probability (namely confidence) and the number of jades of the area belonging to the jade;
secondly, the position and width and height of each jade are determined (coordinates of the center point of the circumscribed rectangle and the width and height of each jade are taken as standards);
③ each jade belongs to probability values of jadeite, Hetian jade, yellow dragon jade, Russian material, seed material, Hsiuyen jade, Duyu, crystal and other categories;
the final classification result of jade detection is mainly to determine which class the jade belongs to according to the probability that the jade belongs to each class. The category with the highest probability can be taken as the category of the jade in the detection result. For example, if the probability that the jade a belongs to [ jade, nephrite, jadeite, yellow dragon jade, russian material, seed material, hsiuyen jade, Duyu, crystal, and others ] is [0.6, 0.05, 0.1, 0.03, 0.02, 0.03, 0.01, 0.09, 0.02, 0.05], the jade a is determined to belong to jadeite. The detection result analysis module 3 also determines whether the confidence exceeds a threshold. And if the confidence coefficient of the detection result does not exceed the threshold N1, performing threshold detection, processing the image and then detecting again. And if the confidence of the detection result exceeds the threshold value N1, the detection result is reliable, and the result output is executed. The confidence threshold N1 is determined according to scene needs, error tolerance rate of jade false detection and other factors, for example, the threshold N1 is 0.6.
And the result output module 4 is mainly used for judging whether the jade in the current picture is completely processed or not and outputting the result. The method comprises the steps of finishing judgment of result detection processing and reporting a result. If the detection result is not processed, executing detection result analysis, and continuously processing the next jade image; and if the detection result is processed, reporting the result. The reported result contains the following information:
firstly, detecting the number, the position and the confidence of the jade in an original image;
② each jade belongs to probability values of jadeite, Hetian jade, yellow dragon jade, Russian material, seed material, Hsiuyen jade, Duyu, crystal and other categories.
The jade classification method comprises the following steps:
s1: acquiring an image to be detected of the jade through image acquisition equipment;
s2: image preprocessing, namely processing the format of the image to achieve the specified size and format;
s3: model detection, namely calculating the number and the class probability value of the jades contained in the image by adopting a pre-trained model;
s4: judging whether a detection result exists or not according to whether the number of the detected jades is not 0 or not, and executing S7 if the detection result exists, namely the number of the detected jades is not less than 1; if no detection result exists, namely the number of the detected jades is 0, executing S5;
s5: whether the threshold value is exceeded or not can be determined according to factors such as scene requirements, computing power performance of a software and hardware platform, fault tolerance rate of jade missing detection and the like, if the number of times of no detection result does not exceed the threshold value N0, S6 is executed, image resolution, brightness, saturation and the like are adjusted, and detection is continued;
if the number of times of no detection result exceeds the threshold value N0, executing S1, wherein the input image does not contain jade, and continuing to process the next image;
s6: image processing, namely processing the format of the image to reach the specified size and format;
s7: analyzing the detection result, namely analyzing the jades one by one, wherein the jade probability, the number of jades, the image positions of the jades, the width and the height and the jade type probability are included;
s8: whether the confidence coefficient exceeds a threshold value or not, if the confidence coefficient of the detection result does not exceed the threshold value N1, executing S5, processing the image and then detecting again; if the confidence of the detection result exceeds the threshold value N1, the detection result is reliable, and S9 is executed;
s9: whether the detection result is processed or not is finished, if the detection result is not processed, S7 is executed, and the next jade image is processed continuously; if the detection result is processed, S10 is executed, and the result is reported.
S10: reporting results, wherein the reported results comprise the number, the position, the confidence coefficient and the jade probability information of the jade.
The jade detection algorithm in the technical scheme of the invention can be replaced by the following target detection algorithm, image segmentation algorithm or classification algorithm, including but not limited to a Yolo series, a MobileNet series, SSD, Pelee, RCNN, FastCNN, FasterRCNN, Mask R-CNN, VGGNet, ResNet, GoogleNet, FCN, SetNet, ReSeg, CRF, MRF and the like, and variant algorithms thereof.
According to the method, jade images are acquired by using image acquisition equipment such as a camera, the input images are subjected to pretreatment such as scaling to a fixed size, features (such as color information, shape information, edge information, texture information and the like) of the images are extracted by using a pre-trained deep learning model algorithm, according to the extracted image features, jade detection and classification are performed by combining the features of various jades of different types learned by the deep learning model algorithm, the number of jades in the current image, the positions of the jades, the types of the jades and the confidence coefficient are determined, and therefore intelligent classification of the jades is achieved.
The method can intelligently and quickly judge whether the jade exists in the input image and the information such as the category of the jade, and greatly improves the efficiency and the accuracy of jade classification.
In the description of the present invention, it should be noted that the terms "first", "second", and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (9)
1. A jade classification method based on image features and a deep learning technology is characterized in that: the jade intelligent classification system comprises an image acquisition module, a jade detection module, a detection result analysis module and a result output module, wherein the image acquisition module acquires jade images through acquisition equipment, preprocesses the input images, extracts features of the images by using a deep learning model algorithm, detects and classifies the jade according to the features of the images and the features of the images, determines the number of the jades in the current images and the positions, the types and the confidence coefficients of the jades, and accordingly realizes intelligent classification of the jades.
2. The method for classifying jade based on image features and deep learning technology according to claim 1, wherein: the acquisition module acquires images by adopting acquisition equipment, wherein the images comprise visible light images and non-visible light images acquired by professional equipment; the jade detection module comprises image preprocessing, model detection, jade number detection and detection result reliability judgment; the detection result analysis module comprises jade quantity analysis, position analysis, category analysis and confidence coefficient analysis, and the result output module comprises detection result judgment completion and information uploading.
3. The method for classifying jade based on image features and deep learning technology according to claim 2, wherein: the acquisition module of the acquisition module comprises but is not limited to a client camera, a professional camera, a scanner, a digital camera, a microscope, an infrared emitter, a flaw detector and the like.
4. The method for classifying jade based on image features and deep learning technology according to claim 2, wherein: the image preprocessing and image processing modes include but are not limited to scaling, translation, transposition, mirroring, rotation, normalization, dimensionality reduction, denoising, equalization, smoothing, and the like.
5. The method for classifying jade based on image features and deep learning technology according to claim 2, wherein: the method of training the model for model detection may use a target detection algorithm, an image segmentation algorithm, or a classification algorithm.
6. The method for classifying jade based on image features and deep learning technology according to claim 2, wherein: the method comprises the following implementation steps:
s1: acquiring an image to be detected of the jade through image acquisition equipment;
s2: image preprocessing, namely processing the format of an image to reach a specified size;
s3: model detection, namely calculating the number and the class probability value of the jades contained in the image by adopting a pre-trained model;
s4: judging whether a detection result exists or not according to whether the number of the detected jades is not 0 or not, and executing S7 if the detection result exists, namely the number of the detected jades is not less than 1; if no detection result exists, namely the number of the detected jades is 0, executing S5;
s5: whether the threshold value is exceeded or not can be determined according to factors such as scene requirements, computing power performance of a software and hardware platform, fault tolerance rate of jade missing detection and the like, if the number of times of no detection result does not exceed the threshold value N0, S6 is executed, image resolution, brightness, saturation and the like are adjusted, and detection is continued;
if the number of times of no detection result exceeds the threshold value N0, executing S1, wherein the input image does not contain jade, and continuing to process the next image;
s6: image processing, namely processing the format of the image to reach the specified size and format;
s7: analyzing the detection result, namely analyzing the jades one by one, wherein the jade probability, the number of jades, the image positions of the jades, the width and the height and the jade type probability are included;
s8: whether the confidence coefficient exceeds a threshold value or not, if the confidence coefficient of the detection result does not exceed the threshold value N1, executing S5, processing the image and then detecting again; if the confidence of the detection result exceeds the threshold value N1, the detection result is reliable, and S9 is executed;
s9: whether the detection result is processed or not is finished, if the detection result is not processed, S7 is executed, and the next jade image is processed continuously; if the detection result is processed, S10 is executed, and the result is reported;
s10: reporting results, wherein the reported results comprise the number, the position, the confidence coefficient and the jade probability information of the jade.
7. The method of claim 6, wherein the jade classification method based on image features and deep learning technique comprises: the jade probability in S7 and S10 is probability value of jadeite, Hetian jade, yellow dragon jade, Russian material, seed material, Xiuyuyan jade, Duyu, crystal, and other types.
8. The method of claim 7, wherein the jade classification method based on image features and deep learning technique comprises: the image formats, files, and resolutions in S1, S2, and S6 are not limited by upper and lower limits of size, nor by the image format.
9. The method of claim 8, wherein the jade classification method based on image features and deep learning technique comprises: the models in S3 include, but are not limited to, the Yolo series, the MobileNet series, SSD, Pelee, RCNN, FastRCNN, FasterRCNN, Mask R-CNN, VGGNet, ResNet, GoogleNet, FCN, SetNet, ReSeg, CRF, MRF, and the like, as well as variant algorithms thereof.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110316436.1A CN113052234A (en) | 2021-03-25 | 2021-03-25 | Jade classification method based on image features and deep learning technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110316436.1A CN113052234A (en) | 2021-03-25 | 2021-03-25 | Jade classification method based on image features and deep learning technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113052234A true CN113052234A (en) | 2021-06-29 |
Family
ID=76515230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110316436.1A Pending CN113052234A (en) | 2021-03-25 | 2021-03-25 | Jade classification method based on image features and deep learning technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113052234A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113486905A (en) * | 2021-07-01 | 2021-10-08 | 深圳星坊科技有限公司 | Jewelry characteristic information acquisition method and device, computer equipment and storage medium |
| CN116994248A (en) * | 2023-09-25 | 2023-11-03 | 支付宝(杭州)信息技术有限公司 | Texture detection processing method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101950359A (en) * | 2010-10-08 | 2011-01-19 | 郝红卫 | Method for recognizing rock type |
| CN107967334A (en) * | 2017-11-30 | 2018-04-27 | 睿视智觉（深圳）算法技术有限公司 | A kind of bullion sorting technique and like product querying method based on image |
| CN108122009A (en) * | 2017-12-25 | 2018-06-05 | 合肥阿巴赛信息科技有限公司 | Jewelry picture classification system and method based on deep learning |
| CN109615024A (en) * | 2018-12-28 | 2019-04-12 | 东北大学 | A kind of Rock Species intelligence Division identification and localization method |
| CN111160320A (en) * | 2020-01-18 | 2020-05-15 | 李瑶 | Image processing method |
-
2021
- 2021-03-25 CN CN202110316436.1A patent/CN113052234A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101950359A (en) * | 2010-10-08 | 2011-01-19 | 郝红卫 | Method for recognizing rock type |
| CN107967334A (en) * | 2017-11-30 | 2018-04-27 | 睿视智觉（深圳）算法技术有限公司 | A kind of bullion sorting technique and like product querying method based on image |
| CN108122009A (en) * | 2017-12-25 | 2018-06-05 | 合肥阿巴赛信息科技有限公司 | Jewelry picture classification system and method based on deep learning |
| CN109615024A (en) * | 2018-12-28 | 2019-04-12 | 东北大学 | A kind of Rock Species intelligence Division identification and localization method |
| CN111160320A (en) * | 2020-01-18 | 2020-05-15 | 李瑶 | Image processing method |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113486905A (en) * | 2021-07-01 | 2021-10-08 | 深圳星坊科技有限公司 | Jewelry characteristic information acquisition method and device, computer equipment and storage medium |
| CN116994248A (en) * | 2023-09-25 | 2023-11-03 | 支付宝(杭州)信息技术有限公司 | Texture detection processing method and device |
| CN116994248B (en) * | 2023-09-25 | 2024-03-15 | 支付宝(杭州)信息技术有限公司 | Texture detection processing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109409355B (en) | Novel transformer nameplate identification method and device | |
| CN103034838B (en) | A kind of special vehicle instrument type identification based on characteristics of image and scaling method | |
| CN106610969A (en) | Multimodal information-based video content auditing system and method | |
| CN109948566B (en) | Double-flow face anti-fraud detection method based on weight fusion and feature selection | |
| CN114549981A (en) | Intelligent inspection pointer type instrument recognition and reading method based on deep learning | |
| CN112149543B (en) | Building dust recognition system and method based on computer vision | |
| CN112907519A (en) | Metal curved surface defect analysis system and method based on deep learning | |
| CN111445459A (en) | Image defect detection method and system based on depth twin network | |
| CN113052234A (en) | Jade classification method based on image features and deep learning technology | |
| WO2023093151A1 (en) | Image screening method and apparatus, electronic device, and storage medium | |
| CN109509188B (en) | Power transmission line typical defect identification method based on HOG characteristics | |
| CN111709305B (en) | Face age identification method based on local image block | |
| CN113537211A (en) | Deep learning license plate frame positioning method based on asymmetric IOU | |
| CN108154496B (en) | Electric equipment appearance change identification method suitable for electric power robot | |
| CN115841633A (en) | Power tower and power line associated correction power tower and power line detection method | |
| JP2006285956A (en) | Red eye detecting method and device, and program | |
| CN110956184A (en) | Abstract diagram direction determination method based on HSI-LBP characteristics | |
| KR20080079798A (en) | Method of face detection and recognition | |
| CN113989196A (en) | Vision-based earphone silica gel gasket appearance defect detection method | |
| CN115908774B (en) | Quality detection method and device for deformed materials based on machine vision | |
| CN112381794A (en) | Printing defect detection method based on deep convolution generation network | |
| CN109829511B (en) | Texture classification-based method for detecting cloud layer area in downward-looking infrared image | |
| CN103871084B (en) | Indigo printing fabric pattern recognition method | |
| CN111652055B (en) | Intelligent switch instrument identification method based on two-stage positioning | |
| CN115187788A (en) | Crop seed automatic counting method based on machine vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210629 |