CN105095851A - Steel coil position identification method - Google Patents
Steel coil position identification method Download PDFInfo
- Publication number
- CN105095851A CN105095851A CN201510095799.1A CN201510095799A CN105095851A CN 105095851 A CN105095851 A CN 105095851A CN 201510095799 A CN201510095799 A CN 201510095799A CN 105095851 A CN105095851 A CN 105095851A
- Authority
- CN
- China
- Prior art keywords
- picture
- coil
- strip
- white point
- circle
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention belongs to the field of freight train cargo image detection and specifically relates to a steel coil position identification method. The method comprises the following steps of: 1, obtaining top views of train compartments, and converting the top views into gray scale images; 2, judging whether the images are disturbed by hard light; 3, identifying positions of compartment walls; 4, determining disturbed images; 5, removing miscellaneous points in the images; and 6, carrying out steel coil position identification. According to the invention, the image identification technology is adopted, image identification is carried out on the top views of the train compartments which are shot by a camera, steel coils in the train compartments can be rapidly identified, the identification speed is relatively high, the accuracy is high, and the time and manpower consumption for steel coil detection is reduced in the freight train steel coil transportation process.
Description
Technical field
The invention belongs to the goods field of image detection of cargo train, specifically a kind of recognition methods of position of steel coil.
Background technology
When train steel coil transportation, in compartment, the position of coil of strip should ensure the middle position in compartment, if coil of strip has been displaced on the limit in compartment in train vibration influence process, may cause damage to coil of strip.Therefore, in the way of steel coil transportation, detecting the position of coil of strip is a very important process.
Along with the high speed development of railway construction in China, the continuous increase of goods station's vehicles while passing, the train turnover time is short, depend merely on manual detection goods difficulty and working strength large, also can grow various potential safety hazard simultaneously, therefore ensure safety of railway traffic, the importance reducing the working strength of Cargo Inspection operator on duty just highlights gradually.
The detection of current train goods takes the photo at the left side of each railway carriage, right side and top mainly through the high-definition camera be arranged on platform, photo is sent to train freight detection place, by manually observing often opening photo, judge in photo, whether compartment goods exists exception.This detection method required time is long, consumes manpower large.
Current information-based fast-developing, mode identification technology development, by carrying out image procossing to the photo of shooting, the method detecting compartment goods exception ought to substitute the method for the artificial judgment exception taken time and effort.
Summary of the invention
The object of the invention is to overcome above-mentioned deficiency, a kind of recognition methods of position of steel coil is provided, decrease in cargo train steel coil transportation process time when checking position of steel coil and manpower consumption.
For realizing above-mentioned technical purpose, scheme provided by the invention is: a kind of recognition methods of position of steel coil, comprises the steps.
Step one, the vertical view of the railway carriage of acquisition, and be translated into gray scale picture.
Step 2, judges whether picture has intense laser interfere, step one is transformed the gray scale picture in the vertical direction obtained and is equally divided into
,
,
three parts,
with
the frequency sampling that two parts get a point according to horizontal, ordinate every ten points obtains
,
the number of two parts point is
with
, then sue for peace respectively
,
total gray scale of fractional-sample point is
with
, then calculate
with
the average gray value that two parts are total
; ?
part also calculates average gray value according to horizontal, that ordinate every ten points get point frequency sampling
, then ask
average gray value with
,
the ratio of part average gray value
if this ratio is greater than 1.35, then judge that this center picture part has intense laser interfere.
Step 3, compartment wall location recognition, obtains by step 2
with
the average gray value that two parts are total
with
the average gray value of part
calculate the average gray value of picture
, according to
calculate binary-state threshold
if: without intense laser interfere,
; If there is intense laser interfere,
, with this binary-state threshold
to picture binaryzation, travel through the two-value picture obtained, calculate the white point number often arranged in two-value picture, judge the position being classified as compartment wall of the point that white point is maximum.
Step 4, judge interference picture, if center picture partial shape is close to the bulk white point region of rectangle, its width is set to
, and
> 500, its length is set to
, and
> 1300, calculates two-value picture white point number and is greater than
the continuous print columns of row
if,
, then judge that time picture is interference picture, namely there is no coil of strip in picture, in calculating
time, get rid of in picture
with
each 300 points of two parts, the interference that after preventing picture binaryzation, the white point of upper and lower compartment wall causes.
Step 5, removes the assorted point in picture, if the profile girth in the region at assorted some place is set to
, and
< 200, gets the profile of picture with Suzuki85 algorithm, then contouring length is less than the boundary rectangle of the profile of 200, and the point in rectangle is assorted point, will mix and a little remove.
Step 6, position of steel coil identification, first judges whether there is annular coil of strip in picture, if having, then exports position of steel coil; If no, then judge whether there is rectangle coil of strip in picture, if having, export position of steel coil; If no, then judge there is no coil of strip in picture.
And the annular coil of strip recognition methods in described step 6 is, by Hall circle transformation, take out the circle in picture, if the center of circle of circle that Hall circle transformation obtains is
, radius is
, the following arbitrary condition of satisfactory foot that Hall circle transformation produces is effective circle:
A) in bianry image, distance is calculated
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round;
B) in bianry image, distance is calculated
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round;
C) in bianry image, distance is calculated
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round;
Wherein,
.
And the rectangle coil of strip recognition methods in described step 6 is, in binary image, calculate the white point number of the maximum row of white point number, be set to
, the height of each rectangle coil of strip is set to L,
If a)
, then coil of strip is not had in picture;
If b)
and
, then
;
If c)
, then
;
Then on bianry image, calculate the white point number that picture often arranges from left to right
,
If a)
and
, then continuous print white point number is judged
and
the number of row
whether reach
,
if reach, then there is a coil of strip herein;
If b)
, then continuous print white point number is judged
the number of row
whether reach
,
if reach, then there are two coil of strips herein;
At the continuous columns of calculating
time, when the number of the row not meeting the condition of continuity is greater than
time,
, judge now discontinuous.
Beneficial effect of the present invention is: the present invention adopts image recognition technology, image recognition is carried out by the picture of overlooking of the railway carriage photographed camera, the coil of strip in railway carriage can be identified rapidly, recognition speed is very fast, accuracy is high, decreases in cargo train steel coil transportation process time when checking position of steel coil and manpower consumption.
Accompanying drawing explanation
Fig. 1 is process flow diagram of the present invention.
Fig. 2 is the original image needing to have in the compartment of identification annular coil of strip.
Fig. 3 is the gray scale picture of the picture having annular coil of strip.
Fig. 4 is the picture after the binaryzation of the picture having annular coil of strip.
Fig. 5 is the position picture of the annular coil of strip detected.
Fig. 6 is the original image needing to have in the compartment of identification rectangle coil of strip.
Fig. 7 is the gray scale picture of the picture having rectangle coil of strip.
Fig. 8 is the picture after the binaryzation of the picture having rectangle coil of strip.
Fig. 9 is the position picture of the rectangle coil of strip detected.
Embodiment
Below in conjunction with drawings and Examples, the invention will be further described.
The present embodiment provides a kind of recognition methods of position of steel coil, as shown in Figure 1, comprises the steps.
1, by being installed on the camera above railway, the vertical view of the railway carriage of acquisition, as Fig. 2 and Fig. 6, and is translated into gray scale picture, as Fig. 3 and Fig. 7.
2, judge whether picture has intense laser interfere.
The intense laser interfere of picture is that the glimmer middle part of the picture caused of the flashlamp of camera in shooting process has an euphotic zone, in order to avoid the erroneous judgement that euphotic zone causes, first judges whether picture has intense laser interfere.
Be equally divided into transforming the gray scale picture in the vertical direction obtained
,
,
three parts,
with
the frequency sampling that two parts get a point according to horizontal, ordinate every ten points obtains
,
the number of two parts point is
with
, then sue for peace respectively
,
total gray scale of fractional-sample point is
with
, then calculate
with
the average gray value that two parts are total
.?
also average gray value is calculated according to horizontal, that ordinate every ten points get point frequency sampling
, ask
average gray value with
,
the ratio of part average gray value
.And take the highlights of picture that obtains at picture
part, by analytical calculation detected picture, obtains the picture of intense laser interfere
all be greater than
, and without the picture of intense laser interfere
all be less than
, therefore get
cut off value
if,
, then this center picture part has intense laser interfere.
3, compartment wall location recognition.
With obtained above
,
part average gray value
with
average gray value
calculate the average gray value of picture
, according to
calculate binary-state threshold
if: without intense laser interfere,
if there is intense laser interfere,
, with this binary-state threshold
to picture binaryzation.Travel through the two-value picture obtained, as Fig. 4 and Fig. 8, calculate the white point number often arranged in two-value picture, the row of the point that white point is maximum are the position of compartment wall.
4, interference picture is judged.
Interference picture be mainly in compartment load tank car or passenger car overlook picture.After tank car and passenger vehicle vertical view binaryzation, center picture part has the bulk white point region of shape close to rectangle, judges this white point region, can judge to disturb picture.The binaryzation picture analyzing tank car and passenger vehicle obtains, and the width of bulk white point rectangle is all greater than 500, is set to
, length is all greater than 1300, is set to
.Therefore, calculate two-value picture white point number be greater than for
the continuous print columns of row
if,
, this picture is the picture of tank car or passenger vehicle, namely disturbs picture, does not have coil of strip in picture.In calculating
time, get rid of in picture
with
each 300 points of two parts, the interference that after preventing picture binaryzation, the white point of upper and lower compartment wall causes.
5, the assorted point in picture is removed.
Then the assorted point of picture is removed.Assorted point is for disperseing point or the point set of distribution in picture after binaryzation, by analysis of binary picture, the girth in the region at assorted some place is all less than 200, is set to
.Get the profile of picture with Suzuki85 algorithm, then contouring length is less than
the boundary rectangle of profile, the point in rectangle is assorted point, will mix and a little remove.
6, position of steel coil identification.
Coil of strip has two states in railway carriage, horizontal positioned and vertically placement.During horizontal positioned, coil of strip shape is close to annulus, and when vertically placing, position of steel coil is close to rectangle.During image recognition, first judge whether there is annular coil of strip in picture, if having, then export position of steel coil, as Fig. 5; If no, then judge whether there is rectangle coil of strip in picture, if having, export position of steel coil, as Fig. 9; If no, then there is no coil of strip in picture.
6.1, annular coil of strip identification.
By Hall circle transformation, take out the circle in picture.Hall circle transformation can produce be not coil of strip interference circle.If the center of circle of the circle that Hall circle transformation obtains is
, radius is
, analyze picture feature and obtain, if interference circle, then bianry image middle distance
length is
the number of white point be not less than round girth
.
So whether be interference circle by following condition judgment circle.
A) in bianry image, distance is calculated
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round.
But the circle obtained due to Hall circle transformation may connect or circumcircle in circular coil of strip, therefore, may filter out effective circle with above-mentioned judgement.
So Rule of judgment will add following condition in addition, to get rid of the situation that a condition filter falls effectively circle.
B) in bianry image, distance is calculated
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round.
C) in bianry image, distance is calculated
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round.
Wherein,
, be the length value that analysis pictorial information obtains.
As long as the arbitrary condition of satisfactory foot above-mentioned a, b, c that Hall circle transformation produces, be effective circle.
Finally, the effective circle detected is marked, is the position of steel coil detected.If the identification of annular coil of strip does not recognize circle, then perform rectangle coil of strip identifying.
6.2, rectangle coil of strip identification.
In binary image, for getting rid of the interference of upper and lower compartment wall, get rid of in picture
with
each 300 points of two parts, then calculate the white point number of the maximum row of white point number, are set to
, the height of each rectangle coil of strip is set to L.By analysis chart picture, obtain following character.
If a)
, in picture, there is no coil of strip.Because the white point sum of each row does not meet the minimum constructive height of coil of strip in picture.
If b)
and
,
.
If c)
,
.
Then on bianry image, calculate the white point number that picture often arranges from left to right
if,
and
, then continuous print white point number is judged
and
the number of row
whether reach
if reach, then there is a coil of strip herein; If
, then continuous print white point number is judged
the number of row
whether reach
if reach, then there are two coil of strips herein.In calculating
time, because discontinuous situation may appear in the white point inside of coil of strip in binary image, therefore when calculating continuation column, only have the number when the row not meeting the condition of continuity to be greater than
time, just say now discontinuous.Wherein, analyze image information to obtain
,
.
Finally, the rectangle steel label detected is gone out, is the position of steel coil detected.
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvement or distortion, these improve or distortion also should be considered as protection scope of the present invention.
Claims (3)
1. a recognition methods for position of steel coil, comprises the steps:
Step one, the vertical view of the railway carriage of acquisition, and be translated into gray scale picture;
Step 2, judges whether picture has intense laser interfere, step one is transformed the gray scale picture in the vertical direction obtained and is equally divided into
,
,
three parts,
with
the frequency sampling that two parts get a point according to horizontal, ordinate every ten points obtains
,
the number of two parts point is
with
, then sue for peace respectively
,
total gray scale of fractional-sample point is
with
, then calculate
with
the average gray value that two parts are total
; ?
part also calculates average gray value according to horizontal, that ordinate every ten points get point frequency sampling
, then ask
average gray value with
,
the ratio of part average gray value
if this ratio is greater than 1.35, then judge that this center picture part has intense laser interfere;
Step 3, compartment wall location recognition, obtains by step 2
with
the average gray value that two parts are total
with
the average gray value of part
calculate the average gray value of picture
, according to
calculate binary-state threshold
if: without intense laser interfere,
; If there is intense laser interfere,
, with this binary-state threshold
to picture binaryzation, travel through the two-value picture obtained, calculate the white point number often arranged in two-value picture, judge the position being classified as compartment wall of the point that white point is maximum;
Step 4, judge interference picture, if center picture partial shape is close to the bulk white point region of rectangle, its width is set to
, and
> 500, its length is set to
, and
> 1300, calculates two-value picture white point number and is greater than
the continuous print columns of row
if,
, then judge that time picture is interference picture, namely there is no coil of strip in picture, in calculating
time, get rid of in picture
with
each 300 points of two parts, the interference that after preventing picture binaryzation, the white point of upper and lower compartment wall causes;
Step 5, removes the assorted point in picture, if the profile girth in the region at assorted some place is set to
, and
< 200, gets the profile of picture with Suzuki85 algorithm, then contouring length is less than the boundary rectangle of the profile of 200, and the point in rectangle is assorted point, will mix and a little remove;
Step 6, position of steel coil identification, first judges whether there is annular coil of strip in picture, if having, then exports position of steel coil; If no, then judge whether there is rectangle coil of strip in picture, if having, export position of steel coil; If no, then judge there is no coil of strip in picture.
2. the recognition methods of a kind of position of steel coil according to claim 1, is characterized in that: the annular coil of strip recognition methods in described step 6 is, by Hall circle transformation, takes out the circle in picture, if the center of circle of circle that Hall circle transformation obtains is
, radius is
, the following arbitrary condition of satisfactory foot that Hall circle transformation produces is effective circle,
In bianry image, calculate distance
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round;
In bianry image, calculate distance
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round;
In bianry image, calculate distance
length is
the number of white point be
, radius is
the girth of circle be
if,
, be then effectively round;
Wherein,
.
3. the recognition methods of a kind of position of steel coil according to claim 1, is characterized in that: the rectangle coil of strip recognition methods in described step 6 is, in binary image, calculates the white point number of the maximum row of white point number, is set to
, the height of each rectangle coil of strip is set to L,
If
, then coil of strip is not had in picture;
If
and
, then
;
If
, then
;
Then on bianry image, calculate the white point number that picture often arranges from left to right
,
If
and
, then continuous print white point number is judged
and
the number of row
whether reach
,
if reach, then there is a coil of strip herein;
If
, then continuous print white point number is judged
the number of row
whether reach
,
if reach, then there are two coil of strips herein;
At the continuous columns of calculating
time, when the number of the row not meeting the condition of continuity is greater than
time,
, judge now discontinuous.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510095799.1A CN105095851B (en) | 2015-03-04 | 2015-03-04 | A kind of position of steel coil recognition methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510095799.1A CN105095851B (en) | 2015-03-04 | 2015-03-04 | A kind of position of steel coil recognition methods |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105095851A true CN105095851A (en) | 2015-11-25 |
CN105095851B CN105095851B (en) | 2018-04-24 |
Family
ID=54576241
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510095799.1A Active CN105095851B (en) | 2015-03-04 | 2015-03-04 | A kind of position of steel coil recognition methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105095851B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105956619A (en) * | 2016-04-27 | 2016-09-21 | 浙江工业大学 | Container lockhole coarse positioning and tracking method |
CN109635797A (en) * | 2018-12-01 | 2019-04-16 | 北京首钢自动化信息技术有限公司 | Coil of strip sequence precise positioning method based on multichip carrier identification technology |
CN109976348A (en) * | 2019-04-11 | 2019-07-05 | 深圳市大富科技股份有限公司 | A kind of vehicle and its motion control method, equipment, storage medium |
CN110113517A (en) * | 2019-05-27 | 2019-08-09 | 杭州亚美利嘉科技有限公司 | Exempt from the camera and self-power supply device of wiring |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101950354A (en) * | 2010-09-03 | 2011-01-19 | 西安大唐电信有限公司 | System and method for recognizing national railway wagon number information image |
CN102768731A (en) * | 2012-06-29 | 2012-11-07 | 陕西省交通规划设计研究院 | Method and system for automatic positioning and identifying target based on high definition video images |
US20120314910A1 (en) * | 2011-06-09 | 2012-12-13 | Carl Zeiss Smt Gmbh | Method and device for determining the position of a first structure relative to a second structure or a part thereof |
CN103077387A (en) * | 2013-02-07 | 2013-05-01 | 东莞中国科学院云计算产业技术创新与育成中心 | Method for automatically detecting carriage of freight train in video |
-
2015
- 2015-03-04 CN CN201510095799.1A patent/CN105095851B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101950354A (en) * | 2010-09-03 | 2011-01-19 | 西安大唐电信有限公司 | System and method for recognizing national railway wagon number information image |
US20120314910A1 (en) * | 2011-06-09 | 2012-12-13 | Carl Zeiss Smt Gmbh | Method and device for determining the position of a first structure relative to a second structure or a part thereof |
CN102768731A (en) * | 2012-06-29 | 2012-11-07 | 陕西省交通规划设计研究院 | Method and system for automatic positioning and identifying target based on high definition video images |
CN103077387A (en) * | 2013-02-07 | 2013-05-01 | 东莞中国科学院云计算产业技术创新与育成中心 | Method for automatically detecting carriage of freight train in video |
Non-Patent Citations (2)
Title |
---|
CAO MAO-YONG 等: "Video events recognition by improved stochastic parsing based on extended stochastic context-free grammar representation", 《JOURNAL OF BEIJING INSTITUTE OF TECHNOLOGY》 * |
张兰: "钢卷空间位置识别技术的研究", 《软件》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105956619A (en) * | 2016-04-27 | 2016-09-21 | 浙江工业大学 | Container lockhole coarse positioning and tracking method |
CN105956619B (en) * | 2016-04-27 | 2019-05-24 | 浙江工业大学 | A kind of container lockhole coarse positioning and tracking |
CN109635797A (en) * | 2018-12-01 | 2019-04-16 | 北京首钢自动化信息技术有限公司 | Coil of strip sequence precise positioning method based on multichip carrier identification technology |
CN109635797B (en) * | 2018-12-01 | 2023-01-13 | 北京首钢自动化信息技术有限公司 | Steel coil sequence accurate positioning method based on multi-carrier identification technology |
CN109976348A (en) * | 2019-04-11 | 2019-07-05 | 深圳市大富科技股份有限公司 | A kind of vehicle and its motion control method, equipment, storage medium |
CN110113517A (en) * | 2019-05-27 | 2019-08-09 | 杭州亚美利嘉科技有限公司 | Exempt from the camera and self-power supply device of wiring |
Also Published As
Publication number | Publication date |
---|---|
CN105095851B (en) | 2018-04-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103745224B (en) | Image-based railway contact net bird-nest abnormal condition detection method | |
CN105095851A (en) | Steel coil position identification method | |
CN111260629A (en) | Pantograph structure abnormity detection algorithm based on image processing | |
WO2017113805A1 (en) | Method and system for identifying train number and model, and safety check method and system | |
CN101893580A (en) | Digital image based detection method of surface flaw of steel rail | |
CN104101600A (en) | Method and apparatus for detecting fine cracks on cross section of continuous casting slab | |
CN113505671B (en) | Machine vision-based carriage congestion degree determination method, system, device and medium | |
CN112508034B (en) | Freight train fault detection method and device and electronic equipment | |
CN111080607B (en) | Rolling bearing oil slinging fault detection method based on image recognition | |
CN103440503A (en) | Vehicle body color detection and identification method | |
CN114022537B (en) | Method for analyzing loading rate and unbalanced loading rate of vehicle in dynamic weighing area | |
CN111523415A (en) | Image-based two-passenger one-dangerous vehicle detection method and device | |
CN111561967A (en) | Real-time online detection method and system for pantograph-catenary operation state | |
CN114463697A (en) | Loading rate calculation method based on image recognition | |
CN113610014A (en) | System and method for detecting freight vehicle with shielding number plate exceeding limit | |
CN110415221B (en) | Automatic detection method for preventing container truck from being lifted based on image feature point matching | |
CN113657265B (en) | Vehicle distance detection method, system, equipment and medium | |
CN111091557B (en) | Method and system for detecting breakage fault of flange of bearing saddle of railway wagon | |
CN109978879B (en) | Box corner in-groove state detection method based on railway wagon loading video monitoring | |
CN111723793A (en) | Real-time rigid contact net positioning point identification method | |
Bong et al. | Car-park occupancy information system | |
CN116503619A (en) | Loading detection method and device for railway wagon, electronic equipment and storage medium | |
CN115359306A (en) | Intelligent identification method and system for high-definition images of railway freight inspection | |
CN205365645U (en) | Train safety check system | |
CN114005097A (en) | Train operation environment real-time detection method and system based on image semantic segmentation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 430070 Hubei province Wuhan city Hongshan District Hongshan town Wenzhi Street No. 16 building 1-7 layer 1 Applicant after: Wuhan Lide Control Technology Co. Address before: 430070 Hubei province Wuhan city Hongshan District Hongshan town Wenzhi Street No. 16 building 1-7 layer 1 Applicant before: WUHAN LIDE SOFTWARE CO., LTD. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |