CN115731493A - Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition - Google Patents
Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition Download PDFInfo
- Publication number
- CN115731493A CN115731493A CN202211424231.6A CN202211424231A CN115731493A CN 115731493 A CN115731493 A CN 115731493A CN 202211424231 A CN202211424231 A CN 202211424231A CN 115731493 A CN115731493 A CN 115731493A
- Authority
- CN
- China
- Prior art keywords
- image
- lens
- physical characteristic
- rainfall
- extracting
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition, which comprises the steps of preprocessing sample data to obtain a binary image; step two, intelligent recognition of raindrops, which comprises the following steps: s2-1, identifying and detecting a binary image based on an image feature algorithm, and extracting raindrop features; s2-2, recognizing time in the image based on OCR, and recognizing the frame number deviation condition; wherein, the identification frame number offset is: the number of frames elapsed from the first frame of the second from the start of analysis to the next second; s2-3, identifying the type of the lens according to the color characteristics of the image; the lens types comprise a blue screen lens, a large lens, a small lens and an undetermined lens; and step three, counting data, associating the result data with the picture processed in the step two, and outputting the precipitation micro physical characteristic parameters. The method and the device are high in identification efficiency and precision, and can facilitate rain drop identification, measurement and analysis and micro physical characteristic parameter extraction of video images acquired by the video sonde.
Description
Technical Field
The application relates to the technical field of meteorological monitoring, in particular to a rainfall micro physical characteristic parameter extraction and analysis method based on video image identification.
Background
The prediction of the rainfall amount has important significance for weather disaster prevention and reduction, and the rainfall amount is generally measured by a rain gauge at the present stage, so that the hysteresis exists, and the rainfall amount information is difficult to obtain in real time or in advance. If the raindrops can be measured in real time, the current rainfall condition can be analyzed and the subsequent rainfall situation can be predicted conveniently in real time, and further more reasonable disaster prevention and reduction measures can be taken.
In a traditional raindrop measurement mode, such as raindrop spectrometers based on optical principles, the inventor thinks that: the observation can only be carried out on the ground, and the change condition of the micro physical characteristics of the precipitation from the ground to the high altitude is difficult to obtain. The video sonde is internally provided with a high-precision camera and can be carried by a sounding balloon to acquire precipitation video image data of a precipitation system from the ground to high altitude. However, the video image information amount is huge, raindrop information cannot be automatically identified, and rainfall micro physical characteristic parameters are extracted, so that the application provides a new technical scheme.
Disclosure of Invention
In order to facilitate the implementation of rainfall analysis prediction work based on raindrop measurement, the application provides a rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition.
The application provides a rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition, which adopts the following technical scheme:
a rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition comprises the following steps:
step one, sample data preprocessing, which comprises the following steps:
s1-1, acquiring raindrop video data acquired by a sounding device, and performing image frame extraction to obtain an image data set;
s1-2, converting the extracted image into a binary picture;
step two, intelligent recognition of raindrops, which comprises the following steps:
s2-1, identifying and detecting a binary image based on an image feature algorithm, and extracting raindrop features;
s2-2, recognizing time in the image based on OCR, and recognizing the frame number deviation condition; wherein, the identification frame number offset is: the number of frames elapsed from the first frame of the number of seconds from the start of analysis to the next second;
s2-3, identifying the type of the lens according to the color characteristics of the image; the lens types comprise a blue screen lens, a large lens, a small lens and an undetermined lens;
step three, data statistics, which comprises the following steps:
s3-1, judging the shape of each raindrop to be circular, square, oval, rectangular, irregular figure, hexagonal or None according to the length-width ratio of the raindrops in the picture;
and S3-2, counting result data of the preorder step, associating the pictures processed in the step two, and outputting the pictures.
Optionally, the S1-2 includes:
a1, performing gray level processing on the image processed by the S1-1 through a weighted average method;
a2, sharpening the gray level image obtained in the step a1 by a gradient method in a differential method to enhance the edge of raindrops in the image;
and a3, carrying out binarization processing on the sharpened picture obtained in the step a 2.
Optionally, the formula of the weighted average method satisfies:
f (i, j) =0.30R (i, j) +0.59G (i, j) +0.11B (i, j), where R (i, j) is the value of the R channel, G (i, j) is the value of the G channel, and B (i, j) is the value of the B channel.
Optionally, the formula of the gradient method satisfies:
wherein the content of the first and second substances,for the magnitude of the gradient, thresh is a custom threshold, L G At a fixed grey level, e.g. L G =255,comprehension of otherwise: with L G Indicating edges, others left the original background value.
Optionally, the S2-1 includes:
b1, performing edge segmentation on the binary image;
b2, performing contour extraction on the picture after the b1 edge segmentation through a canny operator.
Optionally, the S2-1 further includes:
and b3, filtering the raindrops obtained by extracting the b2 by taking a preset length-width ratio as a condition to remove noise which does not meet the condition.
Optionally, the extracting step of the canny operator includes:
generating a Gaussian filter coefficient;
smoothing the original image by using the generated Gaussian filter coefficient;
solving the gradient of the filtered image;
carrying out non-maximum inhibition;
counting a histogram of the image, and judging a threshold value;
searching a boundary starting point by using a function;
and according to the result of the previous step, searching from a pixel point, and searching all boundary points of a boundary with the pixel point as a boundary starting point.
Optionally, the S2-2 includes:
c1, carrying out time frame positioning on the picture obtained after processing in the step b3 based on the coordinate position of time in the sample video used by the user to obtain a time frame;
and c2, performing OCR recognition on the time frame.
Optionally, the OCR recognizing includes: identified by a neural network, and the network structure comprises a neural network CNN, a recurrent neural network RNN and a Loss function CTC Loss.
In summary, the present application includes at least one of the following beneficial technical effects: the video data collected by the sounding equipment are analyzed and processed through an image recognition technology, and the information such as the area, the perimeter, the long axis, the short axis, the length-width ratio and the shape of raindrops is obtained, so that the rainfall tendency and the like can be analyzed based on the information, the implementation cost of rainfall analysis and prediction work is reduced, and the wide implementation is facilitated.
Drawings
FIG. 1 is a schematic flow diagram of the method of the present application;
FIG. 2 is a schematic diagram of the effect of the present application after edge segmentation;
FIG. 3 is a graph of the picture features extracted by the Canny operator of the present application;
FIG. 4 is a pictorial illustration of an input;
fig. 5 is a schematic of an output video picture.
Detailed Description
The present application is described in further detail below with reference to figures 1-5.
The embodiment of the application discloses a precipitation micro physical characteristic parameter extraction and analysis method based on video image recognition.
Referring to fig. 1, the method for extracting and analyzing rainfall micro physical characteristic parameters based on video image identification comprises the following steps:
preprocessing sample data, namely preprocessing raindrop videos shot by sounding equipment (such as a sounding balloon) to obtain sample images to wait for rainfall analysis;
step two, intelligent identification of raindrops;
and step three, data statistics is carried out, and details (documents) of the processed video and other raindrops (obtained by analysis) are output.
Specifically, it comprises:
s1-1, acquiring raindrop video data acquired by a sounding device, and performing image frame extraction (extracting image information of each frame) to obtain an image data set;
and S1-2, converting the extracted image into a binary image.
As to the binarized picture, it includes:
a1, performing gray level processing on the image processed by the S1-1 through a weighted average method; the original image is an RGB three-channel image, and the three components are weighted and averaged by different weights through a weighted averaging method, so that a more reasonable gray image can be obtained.
The formula of the weighted average method satisfies:
F(i,j)=0.30R(i,j)+0.59G(i,j)+0.11B(i,j)
where R (i, j) is the value of the R channel, G (i, j) is the value of the G channel, and B (i, j) is the value of the B channel.
and a2, sharpening the gray-scale image obtained in the step a1 by a gradient method in a differential method to enhance the edge of the raindrop in the image.
The formula of the gradient method satisfies:
for the magnitude of the gradient, thresh is a custom-defined threshold, L G For a fixed grey level, e.g. L G =255,understanding of otherwise: with L G Indicating edges, others left the original background value.
a3, carrying out binarization processing on the sharpened picture obtained in the step a 2; the image is only in black and white colors through binarization processing, and the data volume in the image is greatly reduced through binarization of the image, so that the outline of a target can be highlighted.
After the above steps are completed, the contents of the second step can be carried out; specifically, the method comprises the following steps:
s2-1, identifying and detecting the binary image based on an image feature algorithm, and extracting raindrop features.
The implementation mode comprises the following steps:
b1, performing edge segmentation on the binary image;
it can be understood that the left and right sides of the original picture have two large black edges, and if the two large black edges are not processed, the two large black edges can be mistakenly identified as raindrops, so that the accuracy of raindrop identification is seriously influenced;
the effect after edge splitting is shown in fig. 2.
b2, performing contour extraction on the picture after the edge segmentation of the b1 through a canny operator.
The reason for the above steps is: the canny operator can suppress non-maxima based on the edge gradient direction, while a dual-threshold hysteresis thresholding can be performed.
Extracting a canny operator:
1) Generating a Gaussian filter coefficient;
2) Smoothing the original image by using the generated Gaussian filter coefficient;
3) Solving the gradient of the filtered image;
4) Non-maximum inhibition;
5) Counting a histogram of the image, and judging a threshold value;
6) Searching a boundary starting point by using a function;
7) And according to the result executed in the step 6), starting searching from a pixel point, and searching all boundary points of a boundary with the pixel point as a boundary starting point.
In the above extraction process, a gaussian smoothing function formula is applied:
wherein, the parameter σ represents the width of the gaussian filter, and the larger σ is, the wider the frequency band of the gaussian filter is, and the better the smoothing degree is.
Gradient formula:
0=atan2(G y ,G x )
wherein, G y For each pixel point in the image, the nth derivative, G, in the vertical (y) direction x The nth derivative in the horizontal (x) direction is given to each pixel point in the image.
The Canny operator extracts the picture features as shown in fig. 3.
In consideration of noise problem in the image, in step S2-1 of the method, it further comprises: and b3, filtering the raindrops extracted from the b2 by using a preset length-width ratio as a condition to remove noise which does not meet the condition.
After the steps are completed, the following steps can be executed: s2-2, recognizing time in the image based on OCR, and recognizing the frame number deviation condition; wherein, the deviation of the identification frame number is: the number of frames elapsed from the first frame at which the number of seconds of analysis starts to the next second.
Specifically, the method comprises the following steps:
c1, carrying out time frame positioning on the picture obtained after processing in the step b3 based on the coordinate position of time in the sample video used by the user to obtain a time frame;
because the position of the timeframe location in the video is invariant (the video in this method defaults to temporal attributes), the timeframe can be directly truncated by the coordinate location.
And c2, performing OCR recognition on the time frame.
The OCR recognition uses a deep neural network, and the network structure mainly comprises a convolutional neural network CNN, a recurrent neural network RNN and a CTC Loss function.
The identification process of the neural network comprises the following steps: picture input-text detection-text recognition-text output.
After the steps are completed, the following steps can be executed: s2-3, identifying the type of the lens according to the color characteristics of the image; the lens types comprise a blue screen lens, a large lens, a small lens and an undetermined lens.
The method comprises the steps of obtaining a picture, identifying the type of the picture according to the characteristics of different overall colors, and dividing the picture into four types of lenses, namely a blue screen lens, a large lens, a small lens and an undetermined lens.
When the execution of the above steps is completed, and various basic data are obtained, the third step, data statistics, or in other words: and (4) carrying out statistical screening on the raindrop analysis result, and outputting an analysis video or a detailed document.
Step three, it includes:
s3-1, judging the shape of each raindrop to be circular, square, oval, rectangular, irregular figure, hexagonal or None according to the length-width ratio of the raindrops in the picture;
and S3-2, counting result data of the preorder step, associating the picture processed in the step two, and outputting the picture.
The result data includes: one or more of raindrop quantity, total area, lens type, current time information, current frame information, reference frame offset information, actual offset information, relative frame information, analysis start time, analysis end time, video size information and FPS information, wherein after correlation, the raindrop quantity, the total area, the lens type, the current time information, the current frame information, the reference frame offset information, the actual offset information, the relative frame information, the analysis start time, the analysis end time, the video size information and the FPS information can be displayed on a picture in a picture display stage.
It will be appreciated that the above, when implemented by a corresponding computer program:
1) Selecting whether to preview the processed picture;
2) Whether the analysis result is superimposed on the picture can be selected, and the patterns such as the size of the characters can be adjusted and analyzed;
3) Minimum and maximum raindrop area thresholds can be set, and raindrops exceeding the thresholds will be ignored. The raindrop size can be segmented, raindrop analysis conditions in different ranges are screened and counted, raindrops can be marked by different colors, and patterns such as marking thickness can be adjusted;
4) The starting time and the ending time of the output result can be specified, the output analysis video (such as mp 4) or the detailed information (txt or Excel) can be selected and output according to the requirement, and the output result can be divided according to the time;
5) An aspect ratio threshold may be set beyond which raindrops will be ignored;
6) A maximum same-screen raindrop quantity threshold value can be set, and frames with the total raindrops exceeding the threshold value are treated as invalid frames;
7) A minimum average area threshold value can be set, and the frame with the raindrop average area smaller than the threshold value is treated as an invalid frame;
8) Text information that can be output includes, but is not limited to, the following examples:
referring to fig. 4 and 5, fig. 4 is a pictorial illustration of an input; fig. 5 is a schematic diagram of the video pictures output after the above steps, so that it can be known that the video pictures can help relevant workers to make precipitation analysis trend prediction.
The above embodiments are preferred embodiments of the present application, and the protection scope of the present application is not limited by the above embodiments, so: all equivalent changes made according to the structure, shape and principle of the present application shall be covered by the protection scope of the present application.
Claims (9)
1. A rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition is characterized by comprising the following steps:
step one, preprocessing sample data, which comprises the following steps:
s1-1, acquiring raindrop video data acquired by a sounding device, and performing image frame extraction to obtain an image data set;
s1-2, converting the extracted image into a binary picture;
step two, intelligent recognition of raindrops, which comprises the following steps:
s2-1, identifying and detecting a binary image based on an image feature algorithm, and extracting raindrop features;
s2-2, recognizing time in the image based on OCR, and recognizing the frame number deviation condition; wherein, the identification frame number offset is: the number of frames elapsed from the first frame of the second from the start of analysis to the next second;
s2-3, identifying the type of the lens according to the color characteristics of the image; the lens types comprise a blue screen lens, a large lens, a small lens and an undetermined lens;
step three, data statistics, which comprises the following steps:
s3-1, judging the shape of each raindrop to be circular, square, oval, rectangular, irregular figure, hexagonal or None according to the length-width ratio of the raindrops in the picture;
and S3-2, counting result data of the preorder step, associating the picture processed in the step two, and outputting the picture.
2. The method for extracting and analyzing rainfall micro physical feature parameters based on video image identification as claimed in claim 1, wherein the S1-2 comprises:
a1, performing gray processing on the image processed by the S1-1 through a weighted average method;
a2, sharpening the gray level image obtained in the step a1 by a gradient method in a differential method to enhance the edge of raindrops in the image;
and a3, carrying out binarization processing on the sharpened picture obtained in the step a 2.
3. The method for extracting and analyzing rainfall micro physical characteristic parameters based on video image identification as claimed in claim 2, wherein the formula of the weighted average method satisfies: f (i, j) =0.30R (i, j) +0.59G (i, j) +0.11B (i, j), where R (i, j) is the value of the R channel, G (i, j) is the value of the G channel, and B (i, j) is the value of the B channel.
4. The method for extracting and analyzing rainfall micro physical characteristic parameters based on video image identification as claimed in claim 2, wherein the formula of the gradient method satisfies:
wherein, G [ f (x, y)]For the magnitude of the gradient, thresh is a custom-defined threshold, L G For a fixed grey level, e.g. L G =255,comprehension of otherwise: with L G Indicating edges, others left the original background value.
5. The method for extracting and analyzing rainfall micro physical feature parameters based on video image identification as claimed in claim 2, wherein the S2-1 comprises:
b1, performing edge segmentation on the binary image;
b2, performing contour extraction on the picture after the b1 edge segmentation through a canny operator.
6. The method for extracting and analyzing rainfall micro physical feature parameters based on video image identification as claimed in claim 3, wherein the S2-1 further comprises:
and b3, filtering the raindrops extracted from the b2 by using a preset length-width ratio as a condition to remove noise which does not meet the condition.
7. The method for extracting and analyzing the physical characteristic parameters of the precipitation micro elements based on the video image recognition as claimed in claim 3, wherein: the extraction step of the canny operator comprises the following steps:
generating a Gaussian filter coefficient;
smoothing the original image by using the generated Gaussian filter coefficient;
solving the gradient of the filtered image;
carrying out non-maximum inhibition;
counting a histogram of the image, and judging a threshold value;
searching a boundary starting point by using a function;
and according to the result of the previous step, searching from a pixel point, and searching all boundary points of a boundary with the pixel point as a boundary starting point.
8. The method for extracting and analyzing rainfall micro physical characteristic parameters based on video image recognition according to claim 3, wherein: the S2-2, which comprises:
c1, carrying out time frame positioning on the picture obtained after processing in the step b3 based on the coordinate position of time in the sample video used by the user to obtain a time frame;
and c2, performing OCR recognition on the time frame.
9. The method for extracting and analyzing rainfall micro-physical characteristic parameters based on video image recognition according to claim 8, wherein: the OCR recognition, comprising: identified by a neural network, and the network structure comprises a neural network CNN, a recurrent neural network RNN and a Loss function CTC Loss.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211424231.6A CN115731493A (en) | 2022-11-15 | 2022-11-15 | Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211424231.6A CN115731493A (en) | 2022-11-15 | 2022-11-15 | Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115731493A true CN115731493A (en) | 2023-03-03 |
Family
ID=85296520
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211424231.6A Pending CN115731493A (en) | 2022-11-15 | 2022-11-15 | Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115731493A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117079085A (en) * | 2023-10-16 | 2023-11-17 | 合肥市斑马智行网络技术有限公司 | Training method of raindrop detection model, vehicle control method, device and medium |
CN117214054A (en) * | 2023-11-09 | 2023-12-12 | 中国人民解放军国防科技大学 | Novel video sonde |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103439756A (en) * | 2013-07-31 | 2013-12-11 | 中国人民解放军理工大学 | Natural precipitation particle micro physical characteristic measuring method based on particle forming speed measurement |
CN103593847A (en) * | 2013-11-25 | 2014-02-19 | 中国航天科工集团第三研究院第八三五七研究所 | Raindrop detection analysis method based on machine vision |
CN110059713A (en) * | 2019-03-07 | 2019-07-26 | 中国人民解放军国防科技大学 | Precipitation type identification method based on precipitation particle multi-feature parameters |
CN114463678A (en) * | 2022-01-20 | 2022-05-10 | 邹明忠 | Rainfall type identification method using camera video image |
-
2022
- 2022-11-15 CN CN202211424231.6A patent/CN115731493A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103439756A (en) * | 2013-07-31 | 2013-12-11 | 中国人民解放军理工大学 | Natural precipitation particle micro physical characteristic measuring method based on particle forming speed measurement |
US20150035944A1 (en) * | 2013-07-31 | 2015-02-05 | Pla University Of Science And Technology | Method for Measuring Microphysical Characteristics of Natural Precipitation using Particle Image Velocimetry |
CN103593847A (en) * | 2013-11-25 | 2014-02-19 | 中国航天科工集团第三研究院第八三五七研究所 | Raindrop detection analysis method based on machine vision |
CN110059713A (en) * | 2019-03-07 | 2019-07-26 | 中国人民解放军国防科技大学 | Precipitation type identification method based on precipitation particle multi-feature parameters |
CN114463678A (en) * | 2022-01-20 | 2022-05-10 | 邹明忠 | Rainfall type identification method using camera video image |
Non-Patent Citations (1)
Title |
---|
胡云涛: "基于PMCS的降水微物理特征分析及类型识别方法研究", 《中国优秀硕士学位论文全文数据库基础科学辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117079085A (en) * | 2023-10-16 | 2023-11-17 | 合肥市斑马智行网络技术有限公司 | Training method of raindrop detection model, vehicle control method, device and medium |
CN117079085B (en) * | 2023-10-16 | 2024-01-26 | 合肥市斑马智行网络技术有限公司 | Training method of raindrop detection model, vehicle control method, device and medium |
CN117214054A (en) * | 2023-11-09 | 2023-12-12 | 中国人民解放军国防科技大学 | Novel video sonde |
CN117214054B (en) * | 2023-11-09 | 2024-03-01 | 中国人民解放军国防科技大学 | Novel video sonde |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110334706B (en) | Image target identification method and device | |
CN106651872B (en) | Pavement crack identification method and system based on Prewitt operator | |
WO2018018788A1 (en) | Image recognition-based meter reading apparatus and method thereof | |
CN112149543B (en) | Building dust recognition system and method based on computer vision | |
KR100772506B1 (en) | Method for classification of geological materials using image processing and apparatus thereof | |
CN109409355B (en) | Novel transformer nameplate identification method and device | |
CN115731493A (en) | Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition | |
WO2018145470A1 (en) | Image detection method and device | |
CN107909081B (en) | Method for quickly acquiring and quickly calibrating image data set in deep learning | |
CN101877127B (en) | Image reference-free quality evaluation method and system based on gradient profile | |
CN110210448B (en) | Intelligent face skin aging degree identification and evaluation method | |
CN109918971B (en) | Method and device for detecting number of people in monitoring video | |
CN111060442B (en) | Oil particle detection method based on image processing | |
CN111382704A (en) | Vehicle line-pressing violation judgment method and device based on deep learning and storage medium | |
CN115205223B (en) | Visual inspection method and device for transparent object, computer equipment and medium | |
CN109460722B (en) | Intelligent license plate recognition method | |
CN110175556B (en) | Remote sensing image cloud detection method based on Sobel operator | |
Daniel et al. | Automatic road distress detection and analysis | |
CN106558044B (en) | Method for measuring resolution of image module | |
CN111665199A (en) | Wire and cable color detection and identification method based on machine vision | |
CN111768455A (en) | Image-based wood region and dominant color extraction method | |
CN116542975A (en) | Defect classification method, device, equipment and medium for glass panel | |
CN103295238B (en) | Video real-time location method based on ROI motion detection on Android platform | |
CN113052234A (en) | Jade classification method based on image features and deep learning technology | |
CN114758139B (en) | Method for detecting accumulated water in foundation pit |
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 |