CN107038444A - A kind of image-recognizing method of pointer dial plate - Google Patents
A kind of image-recognizing method of pointer dial plate Download PDFInfo
- Publication number
- CN107038444A CN107038444A CN201610075070.2A CN201610075070A CN107038444A CN 107038444 A CN107038444 A CN 107038444A CN 201610075070 A CN201610075070 A CN 201610075070A CN 107038444 A CN107038444 A CN 107038444A
- Authority
- CN
- China
- Prior art keywords
- pointer
- image
- dial
- pixel
- dial plate
- 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
Classifications
-
- 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
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Input (AREA)
Abstract
The invention discloses a kind of image-recognizing method of pointer dial plate, it includes step：(1) pointer dial plate image is gathered；(2) image is pre-processed；(3) coordinate of each pointer pixel and the left margin point M of dial and right margin point N coordinate are obtained；(4) M and N of dial midpoint S coordinate are determined；(5) it is scanned by abscissa of S abscissa to the direction close to dial, determines dial pixel R；(6) line segment length between left margin point M or right margin point N and midpoint S is set as p, and the line segment length between dial pixel R and midpoint S is q, is determined according to following models from the center of circle O of dial to dial pixel R apart from r：(7) center of circle O of dial coordinate is obtained；(8) inclination angle that all pointer pixels are characterized is obtained；(9) average value at all inclinations angle is asked for as pointer inclination angle；(10) reading of pointer dial plate is obtained.
Description
Technical field
The present invention relates to a kind of recognition methods of dial reading, more particularly to a kind of recognition methods of pointer dial reading.
Background technology
The pointer dial plate image of artificial collection is replaced by machine and it is identified, so that the reading for obtaining pointer dial plate is one of focus for studying at present.Publication No. CN 103955907A, publication date is on July 30th, 2014, entitled " pointer SF6The Chinese patent literature of the method for telemetering of air density meter " discloses a kind of pointer SF6The method of telemetering of air density meter, it includes step：Gather pointer-type SF6The image of air density meter, and by image transmitting to read end；The image collected is changed into black white image；Image specification processing is carried out to black white image, to remove the acute variation of the gradation of image because of caused by illumination acute variation in black white image；Set pointer-type SF6The pointer identification region of air density meter, image binaryzation is carried out to the treated black white image in pointer identification region；Remove the patch noise in pointer identification region；The linear equation corresponding to pointer in pointer identification region is found, and determines the slope of the linear equation；Determine the corresponding reading of pointer direction.
Pointer-type SF disclosed in above-mentioned Chinese patent literature6In the method for telemetering of air density meter, its disclosed described method for finding the linear equation corresponding to pointer is by carrying out Hough transform to the black picture element in pointer identification region.Hough transform is primarily used to recognize the obvious straight line of shape, most long, characteristic straight line, pointer SF especially in image when image is handled6The pointer of air density meter is relatively thick and long, and with dial plate color on the contrary, the straight line that constitutes of the pointer pixel in its image after pretreatment is long and obvious, therefore be relatively adapted to be handled with Hough transform.
Actually, because the specification of different instrument is different, the physical characteristic such as the shape of pointer, thickness, length, color can have different situation in different instrument, if pointer width is thicker, and its color is opposite with dial plate color, the straight line that pointer pixel so in its image after pretreatment is constituted is long and obvious, is easier processing；Conversely, if pointer is thinner, or its color and the color of dial plate are relatively, pointer pixel number so in its image after pretreatment is few and compares concentration, do not allow disposable, or even just have been considered as in preprocessing process the interference noise region of small area and be disposed of in advance.
Fig. 1 shows a kind of black white image for pointer dial plate that the artwork conversion through a kind of pointer dial plate is obtained, and Fig. 2 shows the preprocessed obtained images of Fig. 1.As shown in Figure 1, the pointer of the pointer dial plate is not only very thin, and color is close with as the dial plate color of background, only a bit of is red, so, as shown in Fig. 2 red one section of the pointer of the pointer dial plate has been identified out during including binaryzation and the pretreatment such as anti-phase, and other parts are not all identified.As can be seen here, as this kind of pointers of Fig. 1 are thinner, the recognition point of pointer after the pre-treatment of the color of color and dial plate relatively be：Displayable pixel number is considerably less, and these pixels concentrate on sub-fraction region.Because Hough transform is when image is handled, it is primarily used to recognize the obvious straight line of shape, most long, characteristic straight line especially in image, therefore for this kind of pointer dial plates of Fig. 1, Hough transform is difficult to work, and is not suitable for being used for recognizing pointer position.
In summary, pointer-type SF disclosed in above-mentioned Chinese patent literature6The straight line that pointer pixel of the method for telemetering of air density meter suitable for image after pretreatment is constituted is long and identification of obvious pointer dial reading, and the number for the pointer pixel not being suitable in image after pretreatment is few and compares the identification of the pointer dial reading of concentration.
Number for this kind of pointer pixels of Fig. 2 is few and compares the type of concentration, there is a kind of Recognition of Reading method to be, scanning entirely has the region of pointer pixel, and select two points of wherein lie farthest away, in general be the point of the most upper left corner and last cell, or the most upper right corner and the most lower left corner two kinds of situations of point, be then based on 2 points of the coordinate, its inclination angle is obtained, reading is then drawn.This method is readily applicable to the situation that pointer shows complete display.In this method, the reason for taking farthest two is to reduce error.If two points taken are mutually adjacent, the inclination angle calculated is mainly 45 °, and 90 °, the value that 135 ° of grades are so determined substantially, error is very big.Fig. 3 is the enlarged drawing in Fig. 2 pointer pixel region, as shown in Figure 3, this method scanning for taking two farthest points is used to obtain horizontal span for 5 pixels, vertical span is 9 pixels, it is two farthest points to obtain the point of the most upper left corner and last cell, inclination angle tangent value is tried to achieve, and then tries to achieve total indicator reading.Although the result that this method for taking two farthest points is tried to achieve and actual value are also relatively, but the error that this method is present is not still small, the pixel sample that main cause is just to provide is few, pixel region is small, the value so drawn is easy to that position by two farthest points is still too close to each other and fluctuation, which is presented, to be influenceed, there is the value determined substantially, produce larger error.
The content of the invention
It is an object of the invention to provide a kind of recognition methods of pointer dial reading, the number of pointer pixel of this method suitable for image after pretreatment is few and compares the identification of the pointer dial reading of concentration.
According to foregoing invention purpose, the present invention proposes a kind of recognition methods of pointer dial reading, and it includes step：
(1) pointer dial plate image is gathered；
(2) the pointer dial plate image is pre-processed, to obtain the pixel map with pointer pixel and dial pixel；
(3) coordinate of each pointer pixel and the left margin point M of dial and right margin point N coordinate are obtained；
(4) the left margin point of dial and the midpoint S of right margin point are determined, and obtains midpoint S coordinate；
(5) using midpoint S abscissa as abscissa, it is scanned in the pixel map to the direction close to dial, when scanning is to first pixel, sets it to dial pixel R；
(6) line segment length between left margin point M or right margin point N and midpoint S is set as p, and the line segment length between dial pixel R and midpoint S is q, is determined according to following models from the center of circle O of dial to dial pixel R apart from r：
(7) according to r value, midpoint S coordinate and q, the center of circle O of dial coordinate is obtained；The coordinate of left margin point M and right margin point N in conjunction with dial, obtain the left margin point M and right margin point N of dial inclination angle, and it corresponds to the zero graduation and full scale for representing pointer respectively；
(8) each pointer pixel is connected with dial center of circle O, to obtain the inclination angle that all pointer pixels are characterized；
(9) average value at the inclination angle that all pointer pixels are characterized is asked for, the average value is regard as pointer inclination angle；
(10) inclination angle of left margin point M and right margin point N based on the pointer inclination angle and dial obtain the reading of pointer dial plate.
The design of the recognition methods of pointer dial reading of the present invention is the coordinate by obtaining the center of circle O of dial, so as to obtain the inclination angle that all pointer pixels are characterized, to ask for its average value as pointer inclination angle, and then obtain the reading of pointer dial plate.Specifically：
Pointer dial plate image is obtained first, and it is pre-processed, to obtain the pixel map with pointer pixel and dial pixel；The pretreatment includes correcting the pointer dial plate image rotation so that the ordinate of the left and right boundary point of dial is consistent, and only abscissa is different；Then the pixel map is scanned, obtains the coordinate of each pointer pixel and the left margin point M of dial and right margin point N coordinate；Obtain line segment MN midpoint S abscissa；Afterwards using midpoint S abscissa as abscissa, it is scanned in the pixel map to the direction close to dial, when scanning is to first pixel R, it is a point on dial that pixel R, which is determined, and the left margin point M and right margin point N of this point and dial are on same circular arc；
The inventive method requires that the center of circle that the center of circle of dial and pointer rotate is same point, in fact, if the two centers of circle are misaligned, so pointer necessarily occurs the tip portion of pointer sometimes below dial when rotation, sometimes above dial, changed in uneven, and do not see such situation in practice, therefore the center of circle and the center of circle one of pointer rotation of dial are set to same point.Due to the midpoint that point S is line segment MN, and point R is consistent with point S abscissas, and M, N ordinate are consistent, therefore straight line RS is line segment MN perpendicular bisector.Assuming that the dial center of circle is point O, then line segment OM, OR length are radius r.If line segment MS length is p, line segment RS length is q, then line segment OS length is (r-q).Because straight line RS is line segment MN perpendicular bisector, so triangle OSM is right angled triangle, its length of side meets Pythagorean theorem p2+(r-q)2=r2, deploy bracket and obtain p2-2rq+q2=0, abbreviation is obtained：
Radius r value is thus obtained, with reference to midpoint S coordinate and line segment RS length q, it is possible to obtain dial center of circle O coordinate；The coordinate of left margin point M and right margin point N in conjunction with dial, obtain the left margin point M and right margin point N of dial inclination angle, and it corresponds to the zero graduation and full scale for representing pointer respectively；
Then, each pointer pixel is connected with dial center of circle O, to obtain the inclination angle that all pointer pixels are characterized；Then the average value at the inclination angle that all pointer pixels are characterized is asked for, the average value is regard as pointer inclination angle；The inclination angle for being finally based on the left margin point M and right margin point N of the pointer inclination angle and dial obtains the reading of pointer dial plate.
The number of pointer pixel in image after pretreatment is few and in the case of comparing concentration, because pointer pixel is fluctuated near true pointer position, the average value at the inclination angle characterized with all pointer pixels is as the method at pointer inclination angle so that the fluctuation is largely offset, therefore close to actual value, so the number of pointer pixel of the recognition methods of pointer dial reading of the present invention suitable for image after pretreatment is few and compares the identification of the pointer dial reading of concentration.
Further, in the recognition methods of pointer dial reading of the present invention, the step (2) specifically includes：
The pointer dial plate image collected is changed into black white image；
Image specification processing is carried out to black white image, to remove the acute variation of the gradation of image because of caused by illumination acute variation in black white image；
Target identification region is chosen, image binaryzation is carried out to the treated black white image in target identification region, the target identification region includes pointer and dial.
It is gray level image by the image that collects changes the black white image come, tonal range is 0~255, totally 256 grades in above-mentioned process step；Each gray level for the new images that described image specification processing is obtained will have defined probability density, therefore can remove the acute variation of the gradation of image because of caused by illumination acute variation in the black white image in advance；Region including pointer and dial is set to target identification region, follow-up algorithm process is handled only for the region, and white pixel is all set to beyond the region；The binaryzation of image is exactly that by appropriate threshold value selection, obtaining still can reflect overall and local feature the binary image of image by the gray level image of 256 brightness degrees, the binaryzation of image is conducive to the further processing of image, image is set to become simple, and data volume reduces, the profile of target interested can be highlighted, the pixel that all gray scales are more than or equal to threshold value is judged as belonging to certain objects, its gray value is 255, otherwise these pixel gray values are 0, represent the object area of background or exception, so that whole image shows obvious black and white effect.
Preferably, in the recognition methods of above-mentioned pointer dial reading, described image binaryzation is carried out using adaptive threshold fuzziness method.
The most commonly used dividing method is the threshold method of the maximum equation difference in current image segmentation.When the gray scale difference between the target part and background in image is smaller, i.e. when the double-hump characteristics of grey level histogram is not obvious, just it is less susceptible to determine a suitable threshold value directly with histogram.And be the threshold method in two class regions of selection segmentation by the maximum variance between clusters that Ostu is proposed, when the shape of image grey level histogram has bimodal but either bimodal unobvious with low ebb without obvious low ebb, the result for tending to more be satisfied with using maximum variance Adaptive Thresholding, this method is relatively simple simultaneously, is a kind of wide variety of threshold selection method.Ostu algorithms are based in the present invention, while considering the adaptive response of threshold value selection, image segmentation will be carried out using auto-thresholding algorithm, it is specific as follows：
If the grey level histogram of some image includes two regions, T is the threshold value in two regions of separation.
Region 1 accounts for the area ratio of whole image：
Region 2 accounts for the area ratio of whole image：
The average gray of entire image：
The average gray in region 1：
The average gray in region 2：
Relation of the average gray of entire image between region 1, the average gray value in region 2 be：
μ=μ1θ1+μ2θ2
If the same area has gray scale similar characteristic, and obvious gray difference is shown as between different zones, when two separated by threshold value T interregional gray scale differences are larger, the average gray μ in two regions1、μ2And the average gray μ of entire image difference is also larger, interregional variance is exactly the actual parameter for describing this species diversity, and its expression formula is：
In formula：Illustrate image and split variance between latter two region by threshold value T.Obviously, there are different T values, different interregional variances will be obtained.When divided two interregional variance is up to maximum, it is considered to be the optimal separation state in two regions, threshold value T is thereby determined that：
Other specification need not be manually set with maximum variance decision threshold, be a kind of method for automatically selecting threshold value.
The above-mentioned threshold value based on iteration can be distinguished where the main region of the foreground and background of place's image, and the threshold value obtained by interative computation is good to the segmentation effect of image, can meet the requirement of instrument board binaryzation.
Preferably, in the recognition methods of above-mentioned pointer dial reading, image specification processing is carried out to the black white image using histogram specification method.
Histogrammic regulation is the processing procedure for making histogram be distributed according to specified rule.Statistical relationship between each gray level in gray level image and its pixel count occurred (number of the pixel in the gray level) is drawn, gray level is represented with abscissa, ordinate represents pixel count, then obtains the grey level histogram (characterizing intensity profile probability density) of the gray level image.Image after the specification processing of image histogram has and some standard picture has similar intensity histogram diagram shape, can so overcome the acute variation of illumination.
Further, in the recognition methods of above-mentioned pointer dial reading, the histogram specification method includes step：
Equalization processing is carried out to black white image using histogram equalization, gray level s is obtained；
According to the gray level probability density function of the black white image intentionally got, transforming function transformation function G (z) is obtained；
Gray level s is carried out by inverse transformation z=G using transforming function transformation function G (z)-1(s)。
The specific derivation process of above-mentioned steps is：
Assuming that intensity profile probability density function and the intensity profile probability density function of the image intentionally got that pr (r) and pz (z) represent the original image normalized respectively.Histogram equalization processing is carried out to original image first, that is, seeks the uniform grey level s of original image to the gray level r of original image transforming function transformation function：
Assuming that it is desired that the gray level of obtained image can also utilize following transforming function transformation function equalization
Its inverse process is z=G-1(v), i.e., by wishing that the uniform grey level v of image obtains wishing the gray level z of image.Because to original image and wishing that image has all made equalization processing, ps (s) and pz (z) have same uniform density.If it is possible to the uniform grey level s obtained from original image replaces the uniform grey level v of the hope image in inverse process, its result wishes the gray level z=G-1 (s) of image by the probability density with required hope image.
According to above-mentioned principle, the process of histogram specification processing is obtained：
(a) equalization processing is carried out to original image using histogram equalization；
(b) according to the gray level probability density function pz (z) of the image intentionally got, transforming function transformation function G (z) is obtained；
(c) the gray level s obtained using step (a) makees inverse transformation z=G-1(s)。
The gray level z of the new images obtained in this way will have defined probability density pz (z) in advance.Can be by two transforming function transformation function T (r) and G-1(s) a functional relation is combined into, i.e.,
Z=G-1[T(r)]
Using this formula, it can produce desired grey level distribution from original image.
In addition, working as
G-1[T (r)]=T (r)
When, histogram specification enhancing processing is just reduced to histogram equalization processing.
Further, in the recognition methods of pointer dial reading of the present invention, the pointer dial plate image is gathered using imaging sensor.
Further, in the recognition methods of above-mentioned pointer dial reading, the pointer dial plate image is gathered using CMOS (complementary metal oxide semiconductor) imaging sensor.
The recognition methods of pointer dial reading of the present invention, the number of pointer pixel suitable for image after pretreatment is few and compares the identification of the pointer dial reading of concentration, compensate in the prior art being difficult to corresponding pointer dial plate the defect of reading.
Brief description of the drawings
Fig. 1 is a kind of black white image of pointer dial plate.
Fig. 2 is the preprocessed obtained images of Fig. 1.
Fig. 3 is the enlarged drawing in Fig. 2 pointer pixel region.
Fig. 4 is the schematic diagram of the recognition methods of pointer dial reading of the present invention.
Embodiment
Recognition methods below in conjunction with Figure of description and specific embodiment to pointer dial reading of the present invention is described in further detail.
Fig. 4 is the schematic diagram of the recognition methods of pointer dial reading of the present invention.
With reference to referring to Fig. 4, a kind of embodiment of the recognition methods of pointer dial reading of the invention is, its reading to pointer dial plate is identified, including step：
(1) pointer dial plate image is gathered；
(2) the pointer dial plate image is pre-processed, to obtain the pixel map with pointer pixel and dial pixel；The step is specifically included：
The pointer dial plate image collected is changed into black white image, its specific transform method is, the RGB component for setting a pixel in image is R, G, B respectively, then the computational methods of the gray value of the pixel are G=0.299r+0.587g+0.114b, each pixel is changed in this mode in image, obtains gray level image (black white image)；
Image specification processing is carried out to black white image using histogram specification method, to remove the acute variation of the gradation of image because of caused by illumination acute variation in black white image；The histogram specification method includes step：Equalization processing is carried out to black white image using histogram equalization, original image uniform grey level s is obtained；According to the gray level probability density function of the black white image intentionally got, transforming function transformation function G (z) is obtained；Original image uniform grey level s is carried out to the gray level z=G for the black white image that inverse transformation is intentionally got using transforming function transformation function G (z)-1(s)；
Selection includes the target identification region of pointer and dial, and image binaryzation is carried out using adaptive iteration method thresholding method to the treated black white image in target identification region；
Pretreatment includes correcting the pointer dial plate image rotation so that the ordinate of the left and right boundary point of dial is consistent, and only abscissa is different；
(3) above-mentioned pixel map is scanned, obtains the coordinate of each pointer pixel and the left margin point M of dial coordinate and right margin point N coordinate；
(4) the left margin point M and right margin point N of dial midpoint S are determined, line segment MN midpoint S abscissa is obtained, so as to obtain midpoint S coordinate；
(5) using midpoint S abscissa as abscissa, it is scanned in pixel map to the direction close to dial, when scanning is to first pixel, dial pixel R is set it to, the left margin point M and right margin point N of itself and dial are on same circular arc；
(6) line segment length between left margin point M or right margin point N and midpoint S is set as p, and the line segment length between dial pixel R and midpoint S is q, is determined according to following models from the center of circle O of dial to dial pixel R apart from r, then：
(7) according to r value, midpoint S coordinate and q, the center of circle O of dial coordinate is obtained；The coordinate of left margin point M and right margin point N in conjunction with dial, obtain the left margin point M of dial inclination angle and right margin point N inclination angle, and it corresponds to the zero graduation and full scale for representing pointer respectively；
(8) each pointer pixel is connected with dial center of circle O, to obtain the inclination angle that all pointer pixels are characterized；
(9) average value at the inclination angle that all pointer pixels are characterized is asked for, the average value is regard as pointer inclination angle；
(10) inclination angle of left margin point M and right margin point N based on the pointer inclination angle and dial obtain the reading of pointer dial plate.
The above-mentioned steps of the present embodiment can be realized by automation equipment, wherein step (1) is realized by the IMAQ terminal with cmos image sensor, and step (2)-(10) can be realized by the computer being connected with the IMAQ terminal.
The number of pointer pixel of the present embodiment in image after pretreatment is few and in the case of comparing concentration, because pointer pixel is fluctuated near true pointer position, the average value at the inclination angle characterized with all pointer pixels is as the method at pointer inclination angle so that the fluctuation is largely offset, therefore close to actual value, so the number of pointer pixel of the recognition methods of the pointer dial reading of the present embodiment suitable for image after pretreatment is few and compares the identification of the pointer dial reading of concentration.
Claims (7)
1. a kind of image-recognizing method of pointer dial plate, it is characterised in that including step：
(1) pointer dial plate image is gathered；
(2) the pointer dial plate image is pre-processed, with obtain have pointer pixel and
The pixel map of dial pixel；
(3) coordinate and the left margin point M of dial and the right of each pointer pixel are obtained
Boundary point N coordinate；
(4) the left margin point of dial and the midpoint S of right margin point are determined, and obtains midpoint S's
Coordinate；
(5) using midpoint S abscissa as abscissa, to close to dial in the pixel map
Direction is scanned, and when scanning is to first pixel, sets it to dial pixel R；
(6) line segment length between left margin point M or right margin point N and midpoint S is set as p, scale
Line segment length between disk pixel R and midpoint S is q, is determined according to following models from dial
Center of circle O is to dial pixel R apart from r：
(7) according to r value, midpoint S coordinate and q, the center of circle O of dial coordinate is obtained；
The coordinate of left margin point M and right margin point N in conjunction with dial, obtain the left margin point of dial
M and right margin point N inclination angle, it corresponds to the zero graduation and full scale for representing pointer respectively；
(8) each pointer pixel is connected with dial center of circle O, to obtain all pointer pictures
The inclination angle that vegetarian refreshments is characterized；
(9) average value at the inclination angle that all pointer pixels are characterized is asked for, the average value is made
For pointer inclination angle；
(10) left margin point M's and right margin point N based on the pointer inclination angle and dial inclines
Oblique angle obtains the reading of pointer dial plate.
2. the image-recognizing method of pointer dial plate according to claim 1, it is characterised in that the step
Suddenly (2) are specifically included：
The pointer dial plate image collected is changed into black white image；
Image specification processing is carried out to black white image, to remove in black white image because of illumination acute variation
Caused by gradation of image acute variation；
Target identification region is chosen, the treated black white image in target identification region is schemed
As binaryzation, the target identification region includes pointer and dial.
3. the image-recognizing method of pointer dial plate according to claim 2, it is characterised in that using certainly
Adapt to thresholding method and carry out described image binaryzation.
4. the image-recognizing method of pointer dial plate according to claim 2, it is characterised in that using straight
Side's figure regulationization method carries out image specification processing to the black white image.
5. the image-recognizing method of pointer dial plate according to claim 4, it is characterised in that described straight
Side's figure regulationization method includes step：
Equalization processing is carried out to black white image using histogram equalization, gray level s is obtained；
According to the gray level probability density function of the black white image intentionally got, transforming function transformation function G is obtained
(z)；
Gray level s is carried out by inverse transformation z=G using transforming function transformation function G (z)-1(s)。
6. the image-recognizing method of pointer dial plate according to claim 1, it is characterised in that using figure
As sensor gathers the pointer dial plate image.
7. the image-recognizing method of pointer dial plate according to claim 6, it is characterised in that use CMOS
Imaging sensor gathers the pointer dial plate image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610075070.2A CN107038444A (en) | 2016-02-03 | 2016-02-03 | A kind of image-recognizing method of pointer dial plate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610075070.2A CN107038444A (en) | 2016-02-03 | 2016-02-03 | A kind of image-recognizing method of pointer dial plate |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107038444A true CN107038444A (en) | 2017-08-11 |
Family
ID=59532235
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610075070.2A Pending CN107038444A (en) | 2016-02-03 | 2016-02-03 | A kind of image-recognizing method of pointer dial plate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107038444A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609557A (en) * | 2017-08-24 | 2018-01-19 | 华中科技大学 | A kind of readings of pointer type meters recognition methods |
| CN108491842A (en) * | 2018-03-27 | 2018-09-04 | 康体佳智能科技（深圳）有限公司 | A kind of dial plate identifying system and recognition methods based on neural network |
| CN110009651A (en) * | 2018-12-27 | 2019-07-12 | 浙江大学台州研究院 | For the Anti-interference algorithm of on-line monitoring system to be read under complex environment for instrument visual |
| CN110135420A (en) * | 2019-05-16 | 2019-08-16 | 北京灵汐科技有限公司 | Dial plate state identification method and device, readable storage medium storing program for executing and electronic equipment |
| CN110245624A (en) * | 2019-06-18 | 2019-09-17 | 北京史河科技有限公司 | A kind of non-homogeneous scale recognition methods, device and computer storage medium |
| CN110852333A (en) * | 2019-11-13 | 2020-02-28 | 广东电科院能源技术有限责任公司 | Automatic reading method and device for pointer instrument |
| CN110852954A (en) * | 2019-11-19 | 2020-02-28 | 随锐科技集团股份有限公司 | Image tilt correction method and system for pointer instrument |
| CN111914623A (en) * | 2020-06-17 | 2020-11-10 | 成都飞机工业（集团）有限责任公司 | Method for identifying scale marks of circular-arc scale dial image |
| CN113947720A (en) * | 2021-12-20 | 2022-01-18 | 广东科凯达智能机器人有限公司 | Method for judging working state of density meter |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10288541A (en) * | 1997-04-14 | 1998-10-27 | Meidensha Corp | Objective picture image extracting method of measuring instrument and indicating accuracy inspection method of pointer in this instrument |
| CN103148881A (en) * | 2013-02-06 | 2013-06-12 | 广东电网公司电力科学研究院 | Method and device for automatically reading pointer instrument by video frequency |
| CN103955907A (en) * | 2014-04-17 | 2014-07-30 | 国家电网公司 | Method for telemetering pointer type SF6 gas density meter |
| CN105260709A (en) * | 2015-09-28 | 2016-01-20 | 北京石油化工学院 | Water meter detecting method, apparatus, and system based on image processing |
-
2016
- 2016-02-03 CN CN201610075070.2A patent/CN107038444A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10288541A (en) * | 1997-04-14 | 1998-10-27 | Meidensha Corp | Objective picture image extracting method of measuring instrument and indicating accuracy inspection method of pointer in this instrument |
| CN103148881A (en) * | 2013-02-06 | 2013-06-12 | 广东电网公司电力科学研究院 | Method and device for automatically reading pointer instrument by video frequency |
| CN103955907A (en) * | 2014-04-17 | 2014-07-30 | 国家电网公司 | Method for telemetering pointer type SF6 gas density meter |
| CN105260709A (en) * | 2015-09-28 | 2016-01-20 | 北京石油化工学院 | Water meter detecting method, apparatus, and system based on image processing |
Non-Patent Citations (1)
| Title |
|---|
| 孙凤杰等: "基于图像处理技术的表盘指针角度识别研究", 《中国电机工程学报》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609557B (en) * | 2017-08-24 | 2020-09-08 | 华中科技大学 | Pointer instrument reading identification method |
| CN107609557A (en) * | 2017-08-24 | 2018-01-19 | 华中科技大学 | A kind of readings of pointer type meters recognition methods |
| CN108491842A (en) * | 2018-03-27 | 2018-09-04 | 康体佳智能科技（深圳）有限公司 | A kind of dial plate identifying system and recognition methods based on neural network |
| CN110009651A (en) * | 2018-12-27 | 2019-07-12 | 浙江大学台州研究院 | For the Anti-interference algorithm of on-line monitoring system to be read under complex environment for instrument visual |
| CN110009651B (en) * | 2018-12-27 | 2021-01-05 | 浙江大学台州研究院 | Anti-interference method of instrument visual reading on-line monitoring system |
| CN110135420A (en) * | 2019-05-16 | 2019-08-16 | 北京灵汐科技有限公司 | Dial plate state identification method and device, readable storage medium storing program for executing and electronic equipment |
| CN110135420B (en) * | 2019-05-16 | 2021-11-16 | 北京灵汐科技有限公司 | Dial plate state identification method and device, readable storage medium and electronic equipment |
| CN110245624A (en) * | 2019-06-18 | 2019-09-17 | 北京史河科技有限公司 | A kind of non-homogeneous scale recognition methods, device and computer storage medium |
| CN110852333A (en) * | 2019-11-13 | 2020-02-28 | 广东电科院能源技术有限责任公司 | Automatic reading method and device for pointer instrument |
| CN110852333B (en) * | 2019-11-13 | 2021-07-23 | 南方电网电力科技股份有限公司 | Automatic reading method and device for pointer instrument |
| CN110852954A (en) * | 2019-11-19 | 2020-02-28 | 随锐科技集团股份有限公司 | Image tilt correction method and system for pointer instrument |
| CN110852954B (en) * | 2019-11-19 | 2023-06-23 | 随锐科技集团股份有限公司 | Image inclination correction method and system for pointer instrument |
| CN111914623A (en) * | 2020-06-17 | 2020-11-10 | 成都飞机工业（集团）有限责任公司 | Method for identifying scale marks of circular-arc scale dial image |
| CN113947720A (en) * | 2021-12-20 | 2022-01-18 | 广东科凯达智能机器人有限公司 | Method for judging working state of density meter |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107038444A (en) | A kind of image-recognizing method of pointer dial plate | |
| CN108876743B (en) | Image rapid defogging method, system, terminal and storage medium | |
| CN112949564B (en) | Pointer type instrument automatic reading method based on deep learning | |
| CN115829883B (en) | Surface image denoising method for special-shaped metal structural member | |
| CN107507173B (en) | No-reference definition evaluation method and system for full-slice image | |
| CN109242853B (en) | PCB defect intelligent detection method based on image processing | |
| CN111931776B (en) | Deep learning-based double-pointer instrument reading method | |
| CN110473189B (en) | Text image definition judging method and system | |
| CN109993161B (en) | Text image rotation correction method and system | |
| CN117197140B (en) | Irregular metal buckle forming detection method based on machine vision | |
| CN115330767B (en) | Method for identifying production abnormity of corrosion foil | |
| CN111080661A (en) | Image-based line detection method and device and electronic equipment | |
| CN116485785B (en) | Surface defect detection method for solar cell | |
| CN116630813B (en) | Highway road surface construction quality intelligent detection system | |
| CN116721107B (en) | Intelligent monitoring system for cable production quality | |
| CN112907460B (en) | Remote sensing image enhancement method | |
| CN112200160A (en) | Deep learning-based direct-reading water meter reading identification method | |
| CN117058137B (en) | Visual inspection method for production quality of laminated glass | |
| CN111709964B (en) | PCBA target edge detection method | |
| CN108932516A (en) | It is a kind of rotate text image bearing calibration and device | |
| CN110929562A (en) | Answer sheet identification method based on improved Hough transformation | |
| CN114820631A (en) | Fabric defect detection method capable of resisting texture interference | |
| CN116958125A (en) | Electronic contest host power supply element defect visual detection method based on image processing | |
| CN114140416A (en) | Glass edge detection method and system based on machine vision | |
| CN110930358B (en) | Solar panel image processing method based on self-adaptive algorithm |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170811 |