CN103116753A - Collecting method and device of meter reading - Google Patents
Collecting method and device of meter reading Download PDFInfo
- Publication number
- CN103116753A CN103116753A CN2013100783543A CN201310078354A CN103116753A CN 103116753 A CN103116753 A CN 103116753A CN 2013100783543 A CN2013100783543 A CN 2013100783543A CN 201310078354 A CN201310078354 A CN 201310078354A CN 103116753 A CN103116753 A CN 103116753A
- Authority
- CN
- China
- Prior art keywords
- image
- reading
- matching
- area
- character
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 230000008569 process Effects 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims description 37
- 230000015654 memory Effects 0.000 claims description 37
- 238000005192 partition Methods 0.000 claims description 23
- 238000012545 processing Methods 0.000 claims description 13
- 238000004891 communication Methods 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 6
- 238000012790 confirmation Methods 0.000 claims description 5
- 230000001502 supplementing effect Effects 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 2
- 238000000638 solvent extraction Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 208000033749 Small cell carcinoma of the bladder Diseases 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 201000007710 urinary bladder small cell neuroendocrine carcinoma Diseases 0.000 description 2
- 235000014676 Phragmites communis Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
Images
Landscapes
- Character Input (AREA)
Abstract
The invention discloses a collecting method and a device of meter reading to achieve the purpose of collecting meter reading automatically without using specific meters. The method includes steps of collecting a reading image in the meter reading area; extracting image features of the reading image; matching the image features of the reading image with image features of a reading template, wherein the reading template comprises at least a continuation character template, the continuation character template is an image which displays N characters of the meter reading, N is equal or greater than two, and image features of defined size images are extracted to match while matching with the continuation character template; returning known character result matched with the reading image. No specific meter is used during the whole process of meter reading collecting.
Description
Technical Field
The invention relates to the field of metering, in particular to a meter reading acquisition method and a meter reading acquisition device.
Background
At present, most of the meters used by residents are mechanical meters. The collection of the reading of the mechanical meter is mainly realized by manual household collection. And the manual work is registered to collect the reading, so that a large amount of manual work is consumed, and the collection can not be realized when the owner is not at home.
In the prior art, in order to solve the problem of collecting the meter reading by manual household, the following two automatic collection modes are proposed:
pulse metering type automatic acquisition mode: in the acquisition mode, a metering gauge needs to be disassembled and then is arranged in a magnetic element, a magnetic-sensitive sensor such as a Hall sensor and a reed switch is arranged at a corresponding position outside the gauge, and automatic acquisition is realized by reading signals sensed by the sensor;
direct-reading automatic acquisition mode: the acquisition mode needs to customize the character wheel in the watch, convert the mechanical position of the character wheel into an electric signal and realize automatic acquisition by reading the electric signal.
However, the two automatic reading collecting modes can be realized only by adopting a specific meter, and are not suitable for the existing installed meters, so that the two automatic reading collecting modes cannot be widely applied in practice.
Disclosure of Invention
In view of the above, the main objective of the present invention is to provide a meter reading collecting method and device to achieve the purpose of automatically collecting readings without using a specific meter.
In a first aspect of embodiments of the present invention, there is provided a meter reading acquisition method, which may include:
acquiring a reading image of a reading area of the meter;
extracting image features of the reading image;
matching the image characteristics of the reading image with the image characteristics of a reading template, wherein the reading template at least comprises a continuous character template, the continuous character template is an image presenting N characters of the reading of the meter, and N is more than or equal to 2;
the matching of the image features of the reading image with the continuous character template comprises:
extracting image features of images with preset sizes from the starting position serving as the matching position in the continuous character template towards a preset direction to obtain the image features of the matching position, wherein the matching position in the continuous character template at least comprises the positions of the N characters, and each matching position corresponds to a known character;
matching the image characteristics of the reading image with the image characteristics of the matching position;
taking the known character corresponding to the matching position with the highest matching degree as the known character matched with the reading image;
returning the known characters that match the reading image.
In a second aspect of embodiments of the present invention, there is provided a meter reading acquisition apparatus, which may comprise:
an image collector: the device comprises a reading image used for collecting a reading area of the meter;
an image processor connected to the image collector: the system comprises an image feature extraction module and a template matching module;
wherein the image feature extraction module: the image characteristic used for extracting the reading image;
wherein the template matching module: the reading template matching method is used for matching image features of the reading image with image features of the reading template, wherein the reading template at least comprises a continuous character template, the continuous character template is an image presenting N characters of the reading meter, N is larger than or equal to 2, and the matching of the image features of the reading image with the continuous character template comprises the following steps: extracting image features of images with preset sizes from a starting position serving as a matching position in the continuous character template towards a preset direction to obtain the image features of the matching position, wherein the matching position in the continuous character template at least comprises the positions of the N characters, each matching position corresponds to one known character, the image features of the reading image are matched with the image features of the matching position, and the known character corresponding to the matching position with the highest matching degree is used as the known character matched with the reading image;
a memory coupled to the image processor: the reading template is stored;
a reading feedback connected to the image processor: for returning known characters that match the reading image.
Therefore, the invention has the following beneficial effects:
because the invention collects the reading image of the reading area of the meter, matches the image characteristic of the reading image with the image characteristic of the reading template, and returns the character which is successfully matched to realize the purpose of automatically reading the meter reading, the processing in the whole meter reading collection process has no special requirement on the meter, therefore, the meter reading automatic collection method realized by the invention does not need to adopt a specific meter;
moreover, the invention provides a new continuous character template according to the characteristic that the meter reading has a plurality of different characters, the template is an image presenting N characters of the meter reading, the template has a plurality of matching positions, at least comprises the positions of the N characters, each matching position corresponds to a known character, and when the matching is carried out, the image characteristics of the image at the matching position are extracted from the template to be matched with the reading image, therefore, the continuous character template is equivalent to a template corresponding to a plurality of known characters, and the matching with the reading image containing different characters can be completed by only one continuous character template, thereby achieving the purpose of saving memory and realizing the invention on a low-memory processor.
Drawings
FIG. 1 is a flow chart of a method embodiment provided by the present invention;
FIG. 2 is a schematic diagram of a continuous character template provided by the present invention;
FIG. 3 is a schematic diagram of a rectangular recognition area according to an embodiment of the present invention;
FIG. 4 is a schematic view of a reading image according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a blank identification area according to an embodiment of the present invention;
FIG. 6 is a flow chart of another method embodiment provided by the present invention;
FIG. 7 is a block diagram of an embodiment of an apparatus according to the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
Referring to fig. 1, a flow chart of a meter reading acquisition method provided by the present invention is shown. As shown, the method may include:
s110, collecting a reading image of a reading area of the meter;
s120, extracting image characteristics of the reading image;
s130, matching the image characteristics of the reading image with the image characteristics of a reading template, wherein the reading template at least comprises a continuous character template, the continuous character template is an image presenting N characters of the reading of the meter, and N is more than or equal to 2;
s140, the matching the image features of the reading image with the continuous character template includes:
s141, extracting image features of an image with a preset size from a starting position serving as a matching position in the continuous character template towards a preset direction to obtain the image features of the matching position, wherein the matching position in the continuous character template at least comprises the positions of the N characters, and each matching position corresponds to a known character;
s142, matching the image characteristics of the reading image with the image characteristics of the matching position;
s143, taking the known character corresponding to the matching position with the highest matching degree as the known character matched with the reading image;
and S150, returning known characters matched with the reading image.
By applying the method embodiment, the image collector can be arranged at a proper position above the meter dial plate to collect the reading image of the reading area of the meter, and because the method embodiment collects the reading image of the reading area of the meter, matches the image characteristics of the reading image with the image characteristics of the reading template, and returns the known characters matched with the reading image to realize the purpose of automatically collecting the meter reading, the method has no special requirements on the meter in the whole meter reading collection process, and therefore, the method for automatically collecting the meter reading does not need to adopt a specific meter.
Moreover, because the traditional template is a template corresponding to a known character, the meter has a plurality of different characters, and the last reading of the general meter may rotate continuously in the shooting process to make the reading in the reading image present an intermediate state of two characters, the traditional template matching method needs a very large number of single-character templates and the single-character template representing the intermediate state to complete the matching with the reading image of the meter, and further generates a large number of reading templates to occupy the memory, therefore, in order to reduce the memory occupation, the invention provides a new continuous character template, the continuous character template is different from the single-character template in the prior art, the template is an image presenting N characters of the reading of the meter, and has a plurality of matching positions, at least including the positions of the N characters, each matching position corresponds to a known character, when matching is carried out, the image characteristics of the image at the matching position are extracted from the image and matched with the reading image, and the known character corresponding to the matching position with the highest matching degree is taken as the known character matched with the reading image, so that the continuous character template provided by the invention is equivalent to a template corresponding to a plurality of known characters.
The N characters in the continuous character template may be arranged in any manner, and of course, it is preferable that the N characters of the continuous character template are arranged in a straight line in the image at intervals of a predetermined row of pixels one by one according to a reading continuous change sequence, and this embodiment is particularly suitable for matching a reading image presenting an intermediate state of two characters.
For example, as shown in fig. 2, the continuous character template may include an image representing digital characters of meter readings 0 to 9, wherein the digital characters 0 to 9 are arranged in a direction perpendicular to the horizontal direction in the word wheel rotation sequence at predetermined rows of pixels one by one, wherein the continuous character template may be obtained in advance by:
acquiring a gray histogram of each character of meter reading, for example, acquiring a gray histogram of 10 characters from 0 to 9 of a meter;
normalizing the gray level histograms of the individual characters to a standard size, e.g., 12 × 18;
the gray level histograms of the respective characters are spliced into one image in a direction perpendicular to the horizontal direction by black pixels spaced by a predetermined line one by one in a rotation order, for example, finally into a sample image having a height of 18 × N + K × 9, where N is the number of meter reading characters and K is the number of rows of black pixels of the predetermined line.
It should be noted that, in the present invention, the reading template may also include a single character template, and the matching method may also adopt a traditional method such as a Support Vector Machine (SVM) character classification matching algorithm to perform matching, for which, the present invention is not limited, for example: when the method is implemented, at least the last character recognition area presented in the reading image can be matched with the continuous character template, and other character recognition areas can be matched with the traditional single character template; or, all the character recognition areas in the reading image may be matched with the continuous character template, and the matching may be implemented according to actual needs.
In order to acquire a reading image meeting the quality requirement, before acquiring the reading image of the reading area of the meter, the image acquisition device can also be subjected to related settings through an SCCB protocol, such as: setting the sampling window to a standard size, setting clock division parameters, setting image sharpness, setting output image format, and so forth. Of course, the image collector may also be set by other buses according to specific implementation requirements, and the content of the specific configuration may also be set according to the specific implementation requirements, which is not limited in the present invention.
The return of the known character matching the image of the reading can be returned to the remote control center over the communication bus as needed to remotely acquire the meter reading.
Because the reading image of the reading area of the meter is collected for reading identification, the condition that the reading image does not meet the identification requirement and the matching fails is not eliminated, and the reading image can be sent to a remote control center to facilitate remote manual identification or other processing, and the method is specifically realized as follows:
if the matching fails, the reading image is compressed into a JPEG image and returned to the remote control center through the communication bus.
In the implementation of the invention, if the acquired reading image meets the requirement, the image features of the reading image can be directly extracted after the reading image is acquired. In a preferred embodiment of the present invention, in view of the need of improving the quality of the reading image and reducing the memory, it is proposed to further perform a pre-processing on the acquired reading image to make the quality and size thereof meet the practical implementation needs, for example, the pre-processing may include:
performing noise reduction processing on the reading image by adopting a middle filter and Laplace sharpening operator to improve the quality of the reading image;
a grayscale histogram of the reading image is extracted as an updated reading image to reduce the reading image size.
In the following, some preferred embodiments of the present invention for improving the recognition accuracy and reducing the memory are described in detail:
(embodiment one)
Since the reading image is affected by the actual shooting environment, the reading image may not only present the reading but also present other background interferences, and therefore, it is important to accurately determine the identification area of the reading image to exclude the background interferences. The reading presented by the meter is usually a plurality of digital characters with certain intervals, the present invention improves the recognition accuracy by determining the recognition area of each reading according to the characteristics of the digital characters, for example, in this embodiment, the step S120 of extracting the image characteristics of the reading image may include:
s121, extracting each area which accords with the characteristics of the single character from the reading image to serve as the identification area of each single character;
and S122, extracting the image characteristics of the recognition area of each single character.
In the specific implementation of the present invention, the extraction of the area conforming to the feature of a single character from the reading image may be implemented by extracting an area of a preset fixed position of each character from the reading image, or by extracting a connected area conforming to the feature of a single character from the reading image, which is not limited in the present invention. In the following, a detailed description is given of an implementation method for obtaining the identified region by extracting the connected region, for example, the step S121 may include:
s121a, extracting a connected region in the reading image, for example, the connected region may be extracted by using an algorithm such as a recursive labeling algorithm, a line-by-line labeling algorithm, or a parallel-check structure algorithm, which is not limited in the present invention. In consideration of the requirement of reducing the memory occupation, the invention preferably adopts a 8-neighborhood recursive marking algorithm to realize the extraction of the connected region;
s121b, obtaining a region corresponding to the connected region according to a graph formed by the boundaries of the pixel sets of the connected region;
s121c, judging whether the area corresponding to the connected area accords with the characteristics of a single character or not;
and S121d, if yes, taking the area which accords with the characteristics of the single character as the identification area of the single character.
It should be noted that, the single character feature may be a threshold range of a ratio of a standard single character area to a reading image area, a threshold range of a ratio of a length and a width of a standard single rectangular character, or any other feature that can be determined to be a single character, and the present invention is not limited thereto.
For example, a single character feature may be a threshold range of aspect ratios for a standard single rectangular character, then:
in step S121b, obtaining the area corresponding to the connected area according to the graph formed by the boundary of the pixel set of the connected area may specifically be: obtaining a rectangular area corresponding to the connected area according to a rectangle formed by intersecting rows and columns tangent to the boundary of the pixel set of the connected area;
for example, referring to fig. 3, a connected region 301 and a connected region 302 displaying black are connected regions extracted by using an 8-neighborhood recursive labeling algorithm, a rectangular region formed by intersecting rows and columns tangent to the boundary of the connected region 301 is a rectangular region 310, and a rectangular region formed by intersecting rows and columns tangent to the boundary of the connected region 302 is a rectangular region 320;
the specific implementation of step S121c for determining whether the rectangular region corresponding to the connected region conforms to the feature of a single character may be:
calculating the length-width ratio of the rectangular area corresponding to the connected area;
judging whether the length-width ratio of the rectangular area is within a preset ratio threshold range, for example, the preset ratio threshold range can be set within a range of [0.5,1.5 ];
if yes, the rectangular area corresponding to the connected area accords with the characteristics of a single character.
In fact, due to the influence of the shooting condition, the reading number presented in the acquired reading image may actually be smaller than half of a character, for example, the last reading number presented in the reading image shown in fig. 4 is smaller than half of a character, since in this embodiment, a rectangular area with a length-width ratio within the interval of [0.5,1.5] is used as the recognition area in the reading image, the recognition area in the reading image will not contain the character, and in order not to influence the number of bits of the recognized reading number, a blank recognition area occupancy may be added at this position, so in this embodiment, the following steps may also be included:
calculating the distance between adjacent areas in the area conforming to the single character feature, for example, the rectangular areas conforming to the single character feature may be sorted according to the abscissa of the specific vertex, wherein the specific vertex may be, for example, a vertex whose abscissa and ordinate are the minimum, and the distance between adjacent rectangular areas is calculated by calculating the difference between the abscissas of the specific vertices of the rectangular areas adjacent to the abscissa;
calculating the interval mean value of all the areas which accord with the characteristics of the single character according to the distance of the adjacent areas;
judging whether the distance between a first area and a second area adjacent to the first area in the areas conforming to the characteristics of the single character is more than two times or more than two times of the interval mean value, if so, filling a blank identification area between the first area and the second area, for example, if the distance is more than N times of the interval mean value, filling N-1 blank identification areas in a non-rectangular area between the first rectangular area and the second rectangular area, wherein the length and the width of each blank identification area can be the mean value of the lengths and the widths of all rectangular areas conforming to the characteristics of the single character;
and judging whether the distance between the area which is adjacent to the reading image boundary and accords with the single character feature and the reading image boundary is more than two times of the average value of the interval or not, and if so, supplementing a blank identification area between the area and the reading image boundary.
For example, as shown in the reading image shown in fig. 4, rectangular regions corresponding to the boundaries of the extracted pixel sets of the connected regions are shown as 501 to 504 in fig. 5, and when it is determined that the rectangular region 504 is more than twice the interval mean from the boundaries of the reading image, a blank recognition region 505 is added.
In addition, considering that when the region conforming to the characteristics of a single character is extracted from the reading image in step S120 as the recognition region, a certain error may occur, and particularly for a character presenting an intermediate state between two characters in the reading image, the extracted recognition region may be only a region where one character is located, and the other character may be regarded as interference exclusion, therefore, in order to reduce the recognition region extraction error, the invention proposes to extend the extracted recognition region to reduce the error rate. Specifically, for example, after extracting each recognition area corresponding to a single character feature from the reading image in step S121, before extracting the image feature of each recognition area in step S122, the method may further include:
performing the following expansion steps on the identification area of each single character to obtain M identification areas of the single character;
accordingly, step S122 may be to extract image features of M recognition regions corresponding to each single character;
step S130 may be to match the image features of the M recognition areas corresponding to each single character with the image features of the reading template;
step S140 may return, for each single character, the known character of the M recognition areas that has the highest degree of matching with the image features of the reading template.
The expanding step comprises:
expanding the recognition area by a preset size according to the position of the recognition area of the single character in the reading image, and obtaining the expanded recognition area from the reading image, for example, assuming that the recognition area of a single character is a rectangular area, the vertex coordinates at both ends of a specific diagonal line of the recognition area are (L, T) and (R, B), wherein L and R are horizontal coordinates and T and B are vertical coordinates, according to the position of the recognition area in the reading image, the side length of the rectangular recognition area can be expanded by one half of the original rectangular area, assuming that the positive direction of the horizontal axis is horizontal to the right and the positive direction of the vertical axis is vertical to the down, the vertex coordinates at both ends of the specific diagonal line of the rectangular recognition area obtained after expanding the recognition area are (L-w/4, T-h/2) and (R + w/4, B + h/2), wherein w = R-L is the rectangular width and h = B-T is the rectangular height;
and extracting M recognition areas with corresponding preset character sizes from the enlarged recognition area according to preset M different preset character sizes as M recognition areas of the single character. For example, the preset M preset character sizes are 6 preset sizes of (w, h) (w, h +1) (w-1, h) (w-1, h-1) (w +1, h) (w +1, h +1), and from the enlarged recognition area, 6 recognition areas of corresponding sizes are extracted as the recognition area of the single character.
After the M recognition regions of the corresponding size are extracted, the recognition rate can be improved by normalizing the recognition regions to the standard size and extracting the image features.
(second embodiment)
First, a detailed description is given of a specific implementation of the step S130 of matching the image features of the reading image with the image features of the consecutive character templates, for example, the step S140 may include:
extracting all matching positions in the continuous character template;
extracting image features of the image with the preset size from each matching position as an initial position towards a preset direction to obtain the image features of each matching position;
matching the image characteristics of the reading image with the image characteristics of each matching position respectively;
and taking the known character corresponding to the matching position with the highest matching degree as the character matched with the reading image.
Specifically, the extracting of all the matching positions in the continuous character template may be implemented by gradually shifting from the original position of the continuous character template according to a preset offset, and extracting the position shifted in each step as the matching position until the position is shifted to the end position of the continuous character template. When the last continuous character in the reading image is subjected to matching recognition, the smaller the preset offset is, the higher the recognition accuracy is, and for the recognition of other characters, the preset offset can be preset to be the standard height of a single character.
Secondly, considering that the image features of the reading image may have high matching degrees with the image features of more than one matching position, and the difference of the matching degrees is smaller than a preset matching difference threshold, for example, due to some interference in the image acquisition process and stain on the surface of a meter dial, the difference of the matching degrees of the image features of the identification area in the reading image and the 2 matching positions with the highest and the second highest matching degrees in the continuous character template is very small and smaller than a certain threshold, in order to obtain a more accurate matching result, secondary matching operation can be performed on the matching positions with the similar matching degrees, and the matching position successfully matched with the reading image is selected according to the result of the secondary matching operation.
(third embodiment)
In order to be able to be implemented on a low-memory processor according to the practical requirement of reducing the cost, this embodiment combines the above preferred embodiments to provide the following preferred embodiments of the method of the present invention, as shown in fig. 6, the method of the present invention can be implemented as:
s610, collecting a reading image of a reading area of the meter;
s611, extracting a gray level histogram of the reading image as an updated reading image;
s620, extracting an area which is not extracted yet and accords with the characteristics of a single character from the reading image to serve as a recognition area of the single character;
s621, extracting image features of the recognition area of the single character;
s630, extracting an unextracted matching position in the continuous character template, wherein each matching position corresponds to a known character;
s631, extracting image features of the image with the preset size from the matching position as a starting position towards a preset direction to obtain the image features of the matching position;
s632, matching the image characteristics of the recognition area of the single character with the image characteristics of the matching position to obtain the matching degree of the matching position;
s633, judging whether the last matching position of the continuous character template is extracted or not, and if not, re-entering the step S630;
s634, if yes, comparing the recognition area corresponding to the single character with the matching degree obtained by matching each matching position, and taking the known character corresponding to the matching position with the highest matching degree as the known character which is successfully matched;
s635, judging whether the last area which accords with the characteristics of a single character is extracted;
s636, if not, setting the matching positions of the continuous character templates to be in an unextracted state, and re-entering the step S620;
and S640, if so, returning the known character successfully matched, and ending the process.
In the embodiment, because the identification area which accords with the characteristics of the single character in the reading image is used for identifying the single character, namely, after one character is identified, the next character is identified, the memory occupied by the intermediate data generated in the character identification process can be released before the next character is identified, and the released memory can be reused in the next character identification process, thereby achieving the purpose of reducing the memory occupation.
Moreover, the invention also provides the following memory arrangement method, for example:
the method comprises the following steps that reading images of a reading area of the meter can be collected, wherein the reading images of the reading area of the meter can be collected by 30-240 in size;
the reading image can be transmitted to a first buffer zone with the size of 60 x 240 through a PPI interface;
after the transmission is finished, extracting the longitudinal axis component of the reading image with the size of 30 × 240, sampling every 4 lines to obtain the reading image with the size of 30 × 60, releasing the first cache region, and storing the reading image with the size of 30 × 60 in a first partition of the first cache region;
the reading image stored in the first partition can be processed into a binary image obtained by processing a 30-60 gray histogram, and the binary image is stored in a second partition of the first cache region;
the image of an identification area which is extracted from the reading image and accords with the characteristics of a single character can be stored in the third partition of the first cache area, and the identification area can be further normalized to be 12-18 identification area and then stored in the fourth partition of the first cache area;
images of preset sizes extracted from continuous character templates can be normalized to 12-18 images and stored in a fifth partition of the first cache region, wherein the continuous character templates can be stored in a second cache region;
the obtained matching degree may be stored in the sixth partition of the first cache area.
In addition, a JPEG image prepared for remote manual recognition may be stored in the third buffer, and an internal variable of the processor implementing the above-described method embodiment may be stored in the fourth buffer.
For example: the overall memory layout can be shown in the following table:
therefore, the memory allocation method can be used for effectively storing the intermediate data generated in the processing flow of the method embodiment, reducing memory fragments and memory occupation, and the method can be implemented on a low-memory processor.
It should be noted that the meter according to the present invention may be a character wheel meter, or any other meter, and the present invention is not limited thereto.
Referring to fig. 7, a composition diagram of a meter reading collecting device provided by the present invention is shown. As shown the apparatus may comprise:
the image collector 710: the device comprises a reading image used for collecting a reading area of the meter;
an image processor 720 coupled to the image collector 710: comprises an image feature extraction module 721 and a template matching module 722;
wherein the image feature extraction module 721: the image characteristic used for extracting the reading image;
wherein the template matching module 722: the reading template matching method is used for matching image features of the reading image with image features of the reading template, wherein the reading template at least comprises a continuous character template, the continuous character template is an image presenting N characters of the reading meter, N is larger than or equal to 2, and the matching of the image features of the reading image with the continuous character template comprises the following steps: extracting image features of images with preset sizes from a starting position serving as a matching position in the continuous character template towards a preset direction to obtain the image features of the matching position, wherein the matching position in the continuous character template at least comprises the positions of the N characters, each matching position corresponds to one known character, the image features of the reading image are matched with the image features of the matching position, and the known character corresponding to the matching position with the highest matching degree is used as the known character matched with the reading image;
memory 730 coupled to image processor 720: the reading template is stored;
a reading feedback 740 coupled to the image processor 720: for returning known characters that match the reading image.
The device embodiment is applied, the device can be installed at a proper position above the meter dial, and the purpose of automatically reading the meter reading is realized by acquiring the reading image of the meter reading area, matching the image characteristics of the reading image with the image characteristics of the reading template and returning the known character successfully matched.
Moreover, according to the characteristic that the meter has a plurality of different characters, and the last reading is likely to continuously rotate in the shooting process to enable the reading in the reading image to present an intermediate state of two characters, the invention stores a new continuous character template in the memory, the continuous character template is different from a traditional template corresponding to a known character template, the template has a plurality of matching positions, each matching position corresponds to a known character, when matching is carried out, the image characteristics of the image of the matching position are extracted from the template to be matched with the reading image, the known character corresponding to the matching position with the highest matching degree is taken as the known character matched with the reading image, therefore, the continuous character template is equivalent to a template corresponding to a plurality of known characters, and the matching with the reading image of the meter can be completed by only one continuous character template, thereby achieving the purpose of saving the memory, the above-described embodiments of the apparatus of the present invention may be implemented on a low memory processor.
Preferably, the memory 730: the character storage device is used for storing a continuous character template, and N characters of the continuous character template are arranged along a straight line in an image at intervals of preset rows of pixels one by one according to a reading continuous change sequence.
In order to acquire a reading image meeting the quality requirement, the image processor 720 of the above device embodiment of the present invention may further include a setting module: can be used for carrying out relevant setting on an image collector through an SCCB protocol, such as: setting a sampling window to be a standard size, setting clock frequency division parameters, setting image sharpness, setting an output image format and the like; or the setting module: the method can be used for carrying out noise reduction processing on the reading image by adopting a middle filtering and Laplace sharpening operator so as to improve the quality of the reading image; a grayscale histogram of the reading image is extracted as an updated reading image to reduce the reading image size.
The reading feedback device can be connected with the remote control center through the communication bus to return the known characters matched with the reading image to the remote control center through the communication bus, and can also be used for compressing the reading image into a JPEG format image to return to the remote control center through the communication bus if the matching fails.
In the following, some preferred embodiments of the present invention for improving the recognition accuracy and reducing the memory are described in detail:
(example four)
In this embodiment, the apparatus embodiment of the present invention improves the recognition accuracy by determining the recognition area of each reading, for example:
the image feature extraction module may include a region extraction unit and a region feature extraction unit;
wherein the region extraction unit: the method can be used for extracting each area which accords with the characteristics of the single character from the reading image as the identification area of each single character;
wherein the region feature extraction unit: can be used to extract image features of the recognition regions of individual characters.
The region extraction unit comprises a connected region extraction subunit, a region primary selection subunit and a region confirmation subunit;
wherein the connected region extracting subunit: can be used to extract connected regions in the reading image;
wherein, the sub-unit is selected primarily in the region: the method can be used for obtaining the area corresponding to the connected area according to the graph formed by the boundary of the pixel set of the connected area;
wherein the area confirmation subunit: the method can be used for judging whether the area corresponding to the connected area accords with the characteristics of a single character, and if so, the area which accords with the characteristics of the single character is taken as the identification area of the single character.
Moreover, the single character feature may be a length-width ratio threshold range of a standard single rectangular character, wherein the region primary selection subunit: the method can be used for obtaining a rectangular area corresponding to the connected region according to a rectangle formed by intersecting rows and columns tangent to the boundary of the pixel set of the connected region;
wherein the area confirmation subunit: the method can be used for calculating the length-width ratio of the rectangular area corresponding to the connected area, judging whether the length-width ratio of the rectangular area is within a preset ratio threshold range, and if so, determining that the rectangular area corresponding to the connected area conforms to the characteristics of a single character.
Considering that the reading presented by the acquired reading image may actually be smaller than half a character due to the influence of the shooting condition, so that the recognition area in the reading image will not contain the character, in order to not influence the number of digits of the recognized reading, the area extracting unit of the embodiment of the present invention may further include:
and (3) area complement selection subunit: the distance between the first area and the adjacent second area is two times or more than two times of the interval mean value, if so, a blank identification area is added between the first area and the second area; and judging whether the distance between the area which is adjacent to the reading image boundary and accords with the single character feature and the reading image boundary is more than two times of the average value of the interval or not, and if so, supplementing a blank identification area between the area and the reading image boundary.
Moreover, in order to reduce the recognition area extraction error, the embodiment of the apparatus of the present invention further expands the extracted recognition area to reduce the error rate, for example, the area extraction unit of the embodiment of the apparatus of the present invention may further include an expansion subunit;
the extension subunit: performing the following expansion steps for the recognition area of each single character to obtain M recognition areas of the single character, wherein the expansion steps comprise: according to the position of the recognition area of the single character in the reading image, the recognition area is expanded by preset size adjustment, the expanded recognition area is obtained from the reading image, and according to M different preset character sizes, M recognition areas with corresponding preset character sizes are extracted from the expanded recognition area to serve as M recognition areas of the single character.
(fifth embodiment)
In this embodiment, a template matching module of an embodiment of the apparatus of the present invention is described in detail first, where the template matching module may include a matching position extraction unit, a matching position image extraction unit, a matching position feature extraction unit, and a matching position matching unit;
wherein the matching position extracting unit: can be used to extract all matching positions in the consecutive character templates from the memory;
wherein the matching position image extracting unit: the image feature extraction method can be used for extracting the image features of the image with the preset size from each matching position as a starting position towards a preset direction to obtain the image features of each matching position;
wherein the matching position matching unit: the method can be used for matching the image characteristics of the reading image with the image characteristics of each position respectively, and the known character corresponding to the matching position with the highest matching degree is taken as the character matched with the reading image.
Secondly, considering that the image features of the reading image may have high matching degrees with the image features of more than one matching position, and the difference of the matching degrees is smaller than a preset matching difference threshold, in order to obtain a more accurate matching result, the matching position matching unit of the embodiment of the invention: and the method can also be used for carrying out secondary matching operation on matching positions with similar matching degrees, and selecting the matching position successfully matched with the reading image according to the result of the secondary matching operation.
(sixth embodiment)
In order to be able to be implemented on low-memory processors, according to the practical need of reducing the cost, the present invention, in combination with the above-mentioned embodiments of the apparatus, proposes the following preferred embodiments of the apparatus, for example:
the image feature extraction module of the embodiment of the invention comprises: the method can be used for extracting an area which is not extracted and accords with the characteristics of a single character from the reading image as a recognition area of the single character, and extracting the image characteristics of the recognition area of the single character;
a template matching module: the method can be used for extracting an unextracted matching position in a continuous character template from a memory, wherein each matching position corresponds to a known character, extracting image features of an image with a preset size from the matching position as a starting position towards a preset direction, obtaining the image features of the matching position, matching the image features of the recognition area of the single character with the image features of the matching position, obtaining the matching degree of the matching position, judging whether the last matching position of the continuous character template is extracted, if not, re-triggering the template matching processing module to execute, if so, comparing the matching degree obtained by matching the recognition area of the single character with each matching position, taking the known character corresponding to the matching position with the highest matching degree as a known character which is successfully matched, and judging whether the last area which accords with the single character features is extracted, if not, setting the matching positions of the continuous character templates to be in a non-extraction state, re-triggering the image feature extraction module to execute, and if yes, triggering the reading feedback device to execute.
Furthermore, the memory of the embodiment of the present invention may also have the following memory arrangements:
the image collector according to the embodiment of the apparatus is configured to collect a reading image of 30 × 240 size, and the memory may include a first buffer area and a second buffer area as shown in the following table, where the first buffer area includes a first partition to a sixth partition;
wherein,
the first buffer area is 60 × 240 in size and used for storing a reading image transmitted by the image collector through the PPI interface;
a first partition: the buffer memory is used for storing the reading image with the size of 30 x 60 obtained after sampling every 4 lines by extracting the longitudinal axis component of the reading image after releasing the first buffer memory;
and a second partition: the image processing device is used for storing a binary image obtained by processing the reading image into a 30-60 gray histogram;
and (3) third partitioning: for storing said extracted one of the identification area images;
a fourth section: for storing an image obtained by normalizing the one identified region image to 12 × 18;
a fifth division area: the character template storage device is used for storing an image obtained by normalizing an image with a preset size extracted from a continuous character template to 12-18;
and a sixth partition: the matching degree storage module is used for storing the obtained matching degree;
a second cache area: for storing successive character templates.
Therefore, the arrangement of the memories in the embodiment of the device can effectively store the intermediate data generated in the processing flow, reduce memory fragments and memory occupation, and the embodiment of the device can be implemented on a low-memory processor.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (23)
1. A meter reading acquisition method, comprising:
acquiring a reading image of a reading area of the meter;
extracting image features of the reading image;
matching the image characteristics of the reading image with the image characteristics of a reading template, wherein the reading template at least comprises a continuous character template, the continuous character template is an image presenting N characters of the reading of the meter, and N is more than or equal to 2;
the matching of the image features of the reading image with the continuous character template comprises:
extracting image features of images with preset sizes from the starting position serving as the matching position in the continuous character template towards a preset direction to obtain the image features of the matching position, wherein the matching position in the continuous character template at least comprises the positions of the N characters, and each matching position corresponds to a known character;
matching the image characteristics of the reading image with the image characteristics of the matching position;
taking the known character corresponding to the matching position with the highest matching degree as the known character matched with the reading image;
returning the known characters that match the reading image.
2. The method of claim 1, wherein extracting image features of the reading image comprises:
extracting each area which accords with the characteristics of the single character from the reading image to be used as the identification area of each single character;
image features of the recognition regions of the respective individual characters are extracted.
3. The method according to claim 2, wherein the extracting the region conforming to the feature of the single character from the reading image as the recognition region of the single character comprises:
extracting a connected region in the reading image;
obtaining a region corresponding to the connected region according to a graph formed by the boundary of the pixel set of the connected region;
judging whether the area corresponding to the connected area accords with the characteristics of a single character or not;
if so, the area which accords with the characteristics of the single character is taken as the identification area of the single character.
4. The method according to claim 3, wherein the rectangular area corresponding to the connected region is obtained according to a rectangle formed by intersecting rows and columns tangent to the boundary of the pixel set of the connected region;
the specific implementation of judging whether the rectangular region corresponding to the connected region conforms to the characteristics of a single character is as follows:
calculating the length-width ratio of the rectangular area corresponding to the connected area;
judging whether the length-width ratio of the rectangular area is within a preset ratio threshold range or not;
if yes, the rectangular area corresponding to the connected area accords with the characteristics of a single character.
5. The method of claim 4, wherein the preset scaling threshold range is [0.5,1.5 ].
6. The method of claim 3, further comprising:
calculating the distance between adjacent areas in the area which accords with the characteristics of the single character;
calculating the interval mean value of all the areas which accord with the characteristics of the single character according to the distance of the adjacent areas;
judging whether the distance between a first area and a second area adjacent to the first area in the areas conforming to the characteristics of the single character is more than twice or more than twice of the interval mean value, and if so, filling a blank identification area between the first area and the second area;
and judging whether the distance between the area which is adjacent to the reading image boundary and accords with the single character feature and the reading image boundary is more than two times of the average value of the interval or not, and if so, supplementing a blank identification area between the area and the reading image boundary.
7. The method of claim 2, wherein before extracting the image features of the respective recognition areas, the method further comprises performing the following expansion steps on the recognition area of each single character to obtain M recognition areas of the single character;
the expanding step comprises:
according to the position of the recognition area of the single character in the reading image, expanding the recognition area by a preset size, and obtaining the expanded recognition area from the reading image;
and extracting M recognition areas with corresponding preset character sizes from the enlarged recognition areas according to M different preset character sizes as M recognition areas of the single character.
8. The method of claim 1, wherein matching image features of the reading image with image features of successive character templates comprises:
extracting all matching positions in the continuous character template;
extracting image features of the image with the preset size from each matching position as an initial position towards a preset direction to obtain the image features of each matching position;
matching the image characteristics of the reading image with the image characteristics of each position respectively;
and taking the known character corresponding to the matching position with the highest matching degree as the character matched with the reading image.
9. The method of claim 1,
the step of extracting image features of the reading image comprises:
extracting an area which is not extracted yet and accords with the characteristics of a single character from the reading image as a recognition area of the single character;
extracting image features of the recognition area of the single character;
the step of matching the image features of the reading image with the image features of the reading template comprises:
extracting an unextracted matching position in the continuous character template, wherein each matching position corresponds to a known character;
extracting image features of the image with the preset size from the matching position as an initial position towards a preset direction to obtain the image features of the matching position;
matching the image characteristics of the recognition area of the single character with the image characteristics of the matching position to obtain the matching degree of the matching position;
judging whether the last matching position of the continuous character template is extracted or not;
if not, re-entering the step of matching the image characteristics of the reading image with the image characteristics of the reading template;
if so, comparing the recognition area of the single character with the matching degree obtained by matching each matching position, and taking the known character corresponding to the matching position with the highest matching degree as the known character successfully matched;
judging whether the last region conforming to the characteristics of a single character is extracted;
if not, setting the matching positions of the continuous character templates to be in an unextracted state, and re-entering the step of extracting the image features of the reading image;
if yes, returning the known character which is successfully matched, and ending the process.
10. The method of claim 9,
the reading image of the reading area of the acquisition meter specifically acquires a reading image of 30 x 240;
the reading image is also transmitted to a first buffer zone with the size of 60 x 240 through the PPI interface;
after the transmission is finished, extracting the longitudinal axis component of the reading image, sampling every 4 lines to obtain the reading image with the size of 30 x 60, releasing the first cache region, and storing the reading image with the size of 30 x 60 in a first partition of the first cache region;
the reading image is processed into a 30-60 gray histogram to obtain a binary image, and the binary image is stored in a second partition of the first cache region;
the extracted image of the identification region is stored in a third partition of the first cache region, and the identification region is further normalized into an image of 12 x 18 and then stored in a fourth partition of the first cache region;
images with preset sizes extracted from continuous character templates are normalized to be 12-18 images and stored in a fifth partition of the first cache region, wherein the continuous character templates are stored in a second cache region;
and the obtained matching degree is stored in the sixth partition of the first cache area.
11. The method of claim 1, wherein the returned match result is returned to a remote control center via a communications bus.
12. The method of claim 1, further comprising:
if the matching fails, the reading image is compressed into a JPEG image and returned to the remote control center through the communication bus.
13. A meter reading collection device, comprising:
an image collector: the device comprises a reading image used for collecting a reading area of the meter;
an image processor connected to the image collector: the system comprises an image feature extraction module and a template matching module;
wherein the image feature extraction module: the image characteristic used for extracting the reading image;
wherein the template matching module: the reading template matching method is used for matching image features of the reading image with image features of the reading template, wherein the reading template at least comprises a continuous character template, the continuous character template is an image presenting N characters of the reading meter, N is larger than or equal to 2, and the matching of the image features of the reading image with the continuous character template comprises the following steps: extracting image features of images with preset sizes from a starting position serving as a matching position in the continuous character template towards a preset direction to obtain the image features of the matching position, wherein the matching position in the continuous character template at least comprises the positions of the N characters, each matching position corresponds to one known character, the image features of the reading image are matched with the image features of the matching position, and the known character corresponding to the matching position with the highest matching degree is used as the known character matched with the reading image;
a memory coupled to the image processor: the reading template is stored;
a reading feedback connected to the image processor: for returning known characters that match the reading image.
14. The apparatus of claim 13, wherein the image feature extraction module comprises a region extraction unit and a region feature extraction unit;
wherein the region extraction unit: extracting each area which accords with the characteristics of the single character from the reading image to be used as a recognition area of each single character;
wherein the region feature extraction unit: for extracting image features of the recognition areas of the individual characters.
15. The apparatus of claim 14, wherein the region extracting unit comprises a connected region extracting subunit, a region initially selecting subunit, and a region confirming subunit;
wherein the connected region extracting subunit: the device is used for extracting a connected region in the reading image;
wherein, the sub-unit is selected primarily in the region: the method comprises the steps of obtaining a region corresponding to a connected region according to a graph formed by the boundary of a pixel set of the connected region;
wherein the area confirmation subunit: and the method is used for judging whether the area corresponding to the connected area accords with the characteristics of the single character, and if so, taking the area which accords with the characteristics of the single character as the identification area of the single character.
16. The apparatus of claim 15, wherein the region initial sub-unit: the rectangular area corresponding to the connected region is obtained according to a rectangle formed by intersecting rows and columns tangent to the boundary of the pixel set of the connected region;
wherein the area confirmation subunit: the method is used for calculating the length-width ratio of the rectangular area corresponding to the connected area, judging whether the length-width ratio of the rectangular area is within a preset ratio threshold range, and if so, enabling the rectangular area corresponding to the connected area to accord with the characteristics of a single character.
17. The apparatus of claim 14, wherein the region extracting unit further comprises a region padding sub-unit: the distance between the first area and the adjacent second area is two times or more than two times of the interval mean value, if so, a blank identification area is added between the first area and the second area; and judging whether the distance between the area which is adjacent to the reading image boundary and accords with the single character feature and the reading image boundary is more than two times of the average value of the interval or not, and if so, supplementing a blank identification area between the area and the reading image boundary.
18. The apparatus of claim 14, wherein the region extracting unit further comprises an expanding subunit;
the extension subunit: performing the following expansion steps for the recognition area of each single character to obtain M recognition areas of the single character, wherein the expansion steps comprise: according to the position of the recognition area of the single character in the reading image, the recognition area is expanded by preset size adjustment, the expanded recognition area is obtained from the reading image, and according to M different preset character sizes, M recognition areas with corresponding preset character sizes are extracted from the expanded recognition area to serve as M recognition areas of the single character.
19. The apparatus of claim 13, wherein the template matching module comprises a matching position extraction unit, a matching position image extraction unit, a matching position feature extraction unit, and a matching position matching unit;
wherein the matching position extracting unit: for extracting all matching positions in the consecutive character templates from the memory;
wherein the matching position image extracting unit: the image feature extraction module is used for extracting image features of images with preset sizes from all the matching positions as starting positions towards a preset direction to obtain the image features of all the matching positions;
wherein the matching position matching unit: the character matching method is used for matching the image characteristics of the reading image with the image characteristics of each position respectively, and the known character corresponding to the matching position with the highest matching degree is used as the character matched with the reading image.
20. The apparatus of claim 13,
the image feature extraction module: extracting an area which is not extracted yet and accords with the characteristics of a single character from the reading image to serve as a recognition area of the single character, and extracting the image characteristics of the recognition area of the single character;
the template matching module: extracting an unextracted matching position in a continuous character template from a memory, wherein each matching position corresponds to a known character, extracting image features of an image with a preset size from the matching position as a starting position towards a preset direction, obtaining the image features of the matching position, matching the image features of a recognition area of a single character with the image features of the matching position, obtaining the matching degree of the matching position, judging whether the last matching position of the continuous character template is extracted, if not, re-triggering a template matching processing module to execute, if so, comparing the matching degree obtained by matching the recognition area of the single character with each matching position, taking the known character corresponding to the matching position with the highest matching degree as a known character which is successfully matched, and judging whether the last area which accords with the single character features is extracted, if not, setting the matching positions of the continuous character templates to be in a non-extraction state, re-triggering the image feature extraction module to execute, and if yes, triggering the reading feedback device to execute.
21. The apparatus of claim 20,
the image collector comprises: for acquiring a reading image of 30 x 240 size;
the memory comprises a first cache region and a second cache region, wherein the first cache region comprises a first partition to a sixth partition;
wherein,
the first buffer area is 60 × 240 in size and used for storing a reading image transmitted by the image collector through the PPI interface;
a first partition: the buffer memory is used for storing the reading image with the size of 30 x 60 obtained after sampling every 4 lines by extracting the longitudinal axis component of the reading image after releasing the first buffer memory;
and a second partition: the image processing device is used for storing a binary image obtained by processing the reading image into a 30-60 gray histogram;
and (3) third partitioning: for storing said extracted one of the identification area images;
a fourth section: for storing an image obtained by normalizing the one identified region image to 12 × 18;
a fifth division area: the character template storage device is used for storing an image obtained by normalizing an image with a preset size extracted from a continuous character template to 12-18;
and a sixth partition: the matching degree storage module is used for storing the obtained matching degree;
a second cache area: for storing successive character templates.
22. The apparatus of claim 13 wherein the reading feedback device is connected to the remote control center via a communication bus and returns the known characters matching the image of the reading to the remote control center via the communication bus.
23. The apparatus of claim 13, wherein the reading feedback: and the image matching device is also used for compressing the reading image into an image in JPEG format and returning the image to the remote control center through the communication bus if the matching fails.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310078354.3A CN103116753B (en) | 2013-03-11 | 2013-03-11 | Collecting method and device of meter reading |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310078354.3A CN103116753B (en) | 2013-03-11 | 2013-03-11 | Collecting method and device of meter reading |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103116753A true CN103116753A (en) | 2013-05-22 |
| CN103116753B CN103116753B (en) | 2014-03-12 |
Family
ID=48415125
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310078354.3A Active CN103116753B (en) | 2013-03-11 | 2013-03-11 | Collecting method and device of meter reading |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103116753B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016112669A1 (en) * | 2015-01-15 | 2016-07-21 | 新天科技股份有限公司 | Character information recognition method based on image processing |
| CN112016549A (en) * | 2020-10-19 | 2020-12-01 | 南昌冠东科技有限公司 | Instrument and meter discernment monitoring device |
| CN112818993A (en) * | 2020-03-30 | 2021-05-18 | 深圳友讯达科技股份有限公司 | Character wheel reading meter end identification method and equipment for camera direct-reading meter reader |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510265A (en) * | 2009-03-23 | 2009-08-19 | 浙江大学 | Association moving number character intelligent identification method |
| CN102840948A (en) * | 2012-09-28 | 2012-12-26 | 长庆石油勘探局技术监测中心 | Batch calibration equipment and batch calibration method for intelligent pointer-type pressure gages |
-
2013
- 2013-03-11 CN CN201310078354.3A patent/CN103116753B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510265A (en) * | 2009-03-23 | 2009-08-19 | 浙江大学 | Association moving number character intelligent identification method |
| CN102840948A (en) * | 2012-09-28 | 2012-12-26 | 长庆石油勘探局技术监测中心 | Batch calibration equipment and batch calibration method for intelligent pointer-type pressure gages |
Non-Patent Citations (2)
| Title |
|---|
| 孙鲁涌等: "基于构成模式的煤气表读数字符识别技术研究", 《现代制造工程》, no. 11, 31 December 2007 (2007-12-31), pages 103 - 105 * |
| 王荣等: "在线字轮数字式仪表的自动读取技术", 《计算机应用》, vol. 28, 31 December 2008 (2008-12-31) * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016112669A1 (en) * | 2015-01-15 | 2016-07-21 | 新天科技股份有限公司 | Character information recognition method based on image processing |
| US10311322B2 (en) | 2015-01-15 | 2019-06-04 | Suntront Tech Co., Ltd | Character information recognition method based on image processing |
| CN112818993A (en) * | 2020-03-30 | 2021-05-18 | 深圳友讯达科技股份有限公司 | Character wheel reading meter end identification method and equipment for camera direct-reading meter reader |
| CN112016549A (en) * | 2020-10-19 | 2020-12-01 | 南昌冠东科技有限公司 | Instrument and meter discernment monitoring device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103116753B (en) | 2014-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107944450B (en) | License plate recognition method and device | |
| CN111681256B (en) | Image edge detection method, image edge detection device, computer equipment and readable storage medium | |
| CN108875746B (en) | License plate recognition method, device and system and storage medium | |
| CN110633620B (en) | Pointer instrument scale identification method and electronic equipment | |
| CN103116753B (en) | Collecting method and device of meter reading | |
| CN112001374B (en) | Cloud detection method and device for hyperspectral image | |
| WO2016112669A1 (en) | Character information recognition method based on image processing | |
| CN108550099B (en) | Method and device for removing watermark in image | |
| CN109034170B (en) | Reading method for circular pointer type instrument of switch cabinet detection device | |
| CN115546793B (en) | Automatic reading method and system for water gauge scales and electronic equipment | |
| CN112669323A (en) | Image processing method and related equipment | |
| CN112052702B (en) | Method and device for identifying two-dimensional code | |
| CN110852207A (en) | Blue roof building extraction method based on object-oriented image classification technology | |
| CN109145912B (en) | Automatic digital instrument reading identification method | |
| CN116862910A (en) | Visual detection method based on automatic cutting production | |
| CN115497010A (en) | Deep learning-based geographic information identification method and system | |
| CN101710340B (en) | Method for retrieving similar images | |
| CN106204616B (en) | Method and device for identifying currency value of Iran paper money | |
| CN108734175A (en) | A kind of extracting method of characteristics of image, device and electronic equipment | |
| CN108053409B (en) | Automatic construction method and system for remote sensing image segmentation reference library | |
| CN113393430A (en) | Thermal imaging image enhancement training method and device for fan blade defect detection | |
| CN109697693B (en) | Method for realizing operation based on big data space | |
| JP2018097574A (en) | Recognition device and self position estimation device | |
| CN111369579A (en) | Method for extracting three-dimensional cloud modeling target area contour | |
| CN110837834A (en) | Digital instrument reading method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| C56 | Change in the name or address of the patentee |
Owner name: HANGZHOU SECK INTELLIGENT TECHNOLOGY CO., LTD. Free format text: FORMER NAME: SHANKE ELECTRONIC TECH. DEV. CO., LTD., HANGZHOU |
|
| CP01 | Change in the name or title of a patent holder |
Address after: 310012, D building, No. 391, No. two, six Wen Lu, Zhejiang, Hangzhou, Xihu District, China Patentee after: HANGZHOU SHANKE INTELLIGENT TECHNOLOGY CO., LTD. Address before: 310012, D building, No. 391, No. two, six Wen Lu, Zhejiang, Hangzhou, Xihu District, China Patentee before: Hangzhou Seck Electronic Technology Development Co., Ltd. |