CN106682664A - Water meter disc area detection method based on full convolution recurrent neural network - Google Patents
Water meter disc area detection method based on full convolution recurrent neural network Download PDFInfo
- Publication number
- CN106682664A CN106682664A CN201611114543.1A CN201611114543A CN106682664A CN 106682664 A CN106682664 A CN 106682664A CN 201611114543 A CN201611114543 A CN 201611114543A CN 106682664 A CN106682664 A CN 106682664A
- Authority
- CN
- China
- Prior art keywords
- water meter
- neural network
- recurrent neural
- full convolution
- disc area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a water meter disc area detection method based on full convolution recurrent neural network. The method comprises the following steps: obtaining the image of a water meter; marking the external rectangular frame at the water meter disc area on the image of the water meter; obtaining the marking information of the external rectangular frame at the water meter disc area; constructing a full convolution recurrent neural network; extracting the multi-channel characteristic image of the water meter image; using a sliding window to scan the multi-channel characteristic image; screening a candidate window for the meter disc area; extracting the corresponding position characteristics of the candidate window for the meter disc area; obtaining a final target detection result; and using the losses of the candidate window for the meter disc area and the final target to update the parameters for the full convolution recurrent neural network. According to the invention, the full convolution recurrent neural network in deep learning is utilized to automatically extract the characteristics of the water meter disc, which solves the problem of detecting a water meter disc area under a complicated environment and further inputs the identified position of the disc as identified by the water meter. This manner greatly increases the identification rate of a water meter.
Description
Technical field
The present invention relates to machine learning and computer vision field, more particularly to the water based on full convolution recurrent neural network
Table disc area detection method.
Background technology
In recent years, with the development of artificial intelligence, solve the problems, such as that conventional machines study runs into into using deep learning
For focus, the water meter disc area detection based on computer vision is exactly an important application in computer vision, and it can be with
Whether water meter disk is included in one pictures of correct identification, be to improve meter reading discrimination to lay a solid foundation, entered
And automatic identification water meter, replace existing artificial meter reading mode.
The problem of the primary solution of water meter disc area detection is exactly the detection of border circular areas, and the method for current main flow mainly has
Hough transformation and the optimization method based on the method.But these methods do not tackle the problem at its root, to various complexity
Illumination, deformation under scene, the condition adaptability such as to block bad.
The content of the invention
To overcome the deficiencies in the prior art, the present invention to propose that the water meter disc area based on full convolution recurrent neural network is examined
Survey method.
The technical scheme is that what is be achieved in that, the water meter disc area based on full convolution recurrent neural network is detected
Method, including step
S1：Water meter image is obtained, the water meter disc area external world rectangle frame on the water meter image is marked, and obtains described
The markup information of water meter disc area external world rectangle frame；
S2：Full convolution recurrent neural network is built, using the full convolution recurrent neural network water meter image is extracted
Multi-channel feature figure；
S3：The multi-channel feature figure is scanned using sliding window, preliminary screening goes out dial plate region candidate window；
S4：The relevant position feature of the dial plate region candidate window position is extracted, final target detection result is obtained；
S5：Using the loss of dial plate region candidate window and final goal loss, the ginseng of the full convolution recurrent neural net is updated
Number.
Further, step S1 includes step
S11：The water meter image pattern in multiple actual scenes is gathered by photographic head, the water meter image pattern includes many
Under planting illumination condition, different visual angles, different type, the different extent of damages, the water meter image pattern of different rotary angle；
S12：The region of water meter disk in the water meter image pattern is labeled, including water meter disc area external world square
Shape frame four apex region coordinate positions (x1, y1), (x2, y2), (x3, y3), (x4, y4).
Further, step S2 includes step
S21：Full convolution recurrent neural network is built, the full convolution recurrent neural network includes multiple convolutional layers and pond
The cascade of layer, is input into as triple channel RGB image, is output as multichannel characteristic pattern；
S22：By error back propagation and stochastic gradient descent method, the parameter of the full convolution recurrent neural network is entered
Row optimization updates.
Further, step S3 includes step
S31：Sliding window scanning is carried out to the multi-channel feature figure, and multi-channel feature figure in sliding window is carried out
Feature Fusion；
S32：The full Connection Neural Network of multiple multilamellars is built, the full Connection Neural Network of each multilamellar is each responsible for different scale
The detection of lower target and positioning.
Further, step S4 includes step
According to the dial plate region candidate window, the feature of relevant position is extracted on the multi-channel feature figure, and carried out
The spatial pyramid pond of sizing, obtains characteristic vector, and characteristic vector obtains target through grader and after returning device calculating
Significance and rectangle frame parameter, the target to detecting carry out it is non-maximization suppress, obtain detect target.
Further, step S5 also includes step
S51：Water meter image in step S1 is replaced with into water meter image to be tested；
S52：Repeat step S2-S4, obtains final target detection result；
S53：Non- maximization is carried out to the final target detection result to suppress, and obtains final target detection result.
The beneficial effects of the present invention is, compared with prior art, the present invention is using the full convolution recurrence in deep learning
Neutral net, automatically extracts water meter disk feature, solves the problems, such as water meter disc area detection under complex background, will identify that
The position of disk is further used as the input of meter reading identification, substantially increases the discrimination of meter reading identification.
Description of the drawings
Fig. 1 is water meter disc area detection method flow chart of the present invention based on full convolution recurrent neural network；
Fig. 2 is the full convolution Recursive Neural Network Structure schematic diagram of the present invention；
Fig. 3 is a water meter image pattern schematic diagram；
Fig. 4 is a water meter disc area testing result schematic diagram.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
Fig. 1 is referred to, water meter disc area detection method of the present invention based on full convolution recurrent neural network includes step
S1：The markup information of water meter image and water meter disc area external world rectangle is obtained, it is big by the collection of RGB photographic head
Water meter image pattern (as shown in Figure 3) in amount actual scene, including under various illumination conditions, different visual angles, different type, no
The same extent of damage, the water meter image pattern of different rotary angle, to ensure the multiformity of sample, lifts water meter region detection
Can, artificial mark, including the rectangle of water meter disc area are carried out to the meter reading region in the water meter image pattern of acquisition
Frame four apex region coordinate positions (x1, y1), (x2, y2), (x3, y3), (x4, y4)；
S2：Multi-channel feature figure is extracted using full convolution recurrent neural network, a full convolution recurrent neural network is designed
(as shown in Figure 2), the convolutional neural networks include the cascade of multiple convolutional layers and pond layer so that the depth convolutional neural networks
It is input into as triple channel RGB image, is output as multichannel characteristic pattern.Its optimization method is the error calculated using loss function
The weighted sum of calculation error：
L=LS3+λ×LS4
By error back propagation and stochastic gradient descent method, the parameter of full convolution recurrent neural network is optimized more
Newly.
S3：Sliding window scanning is carried out to the multi-channel feature figure, target area candidate's window is obtained, specific implementation step is such as
Under：
S31：The multi-channel feature figure that image pattern in S2 is obtained after the calculating of full convolution recurrent neural network is carried out
Sliding window is scanned, and multi-channel feature figure in sliding window is carried out into Feature Fusion；
S32：Input is characterized as with S31 gained, the full Connection Neural Network of multiple multilamellars is designed, under being each responsible for different scale
The detection of target and positioning.
As Overlap ＞ 0.7, this feature as positive sample feature, is returned device with the external horizontal square by the grader
Center, the length and width of shape frame is used as regressive object；
As Overlap ＜ 0.3, this feature as negative sample feature, is returned device not calculation error by the grader；
When 0.7 >=Overlap >=0.3, the grader and device not calculation error is returned；
S33：Using grader be output as Sigmoid functions：
Grader loss function is cross entropy loss function：
The recurrence device loss function for adopting is for Euclidean distance loss function：
S34：According to candidate's window significance of grader output, target area candidate window of the probability more than 0.5 is filtered out, and
Target rectangle frame parameter according to device output is returned carries out non-maximization to the candidate frame for being filtered out and suppresses, and its specific practice is such as
Under：Only retain confidence level highest result in target frame of the Duplication more than 0.5.
S4：Extract relevant position feature and obtain final target detection result according to target area candidate's window position, specifically
Implementation steps are as follows：
S41：According to S34 gained target candidate windows, the feature of relevant position is extracted on multi-channel feature figure, and carry out determining
The spatial pyramid pond of size, obtains characteristic vector；
S42：Characteristic vector is through grader and returns significance and rectangle frame that target is obtained after device is calculated, grader
Device loss function is identical with described in S33 with returning；
S43：Target to detecting carries out non-maximization and suppresses, and obtains detecting target.
S5：Lost into line parameter using the loss of candidate's window and final goal and updated.
When water meter image measurement is carried out, the data of step S1 are replaced with into test data, sequentially pass through S2, S3, S4 step
After rapid, object detection results carried out with non-maximization suppression and obtains final target detection result (as shown in Figure 4).
The above is the preferred embodiment of the present invention, it is noted that for those skilled in the art
For, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (6)
1. the water meter disc area detection method of full convolution recurrent neural network is based on, it is characterised in that including step
S1：Water meter image is obtained, the water meter disc area external world rectangle frame on the water meter image is marked, and obtains the water meter
The markup information of disc area external world rectangle frame；
S2：Full convolution recurrent neural network is built, using the full convolution recurrent neural network many of the water meter image are extracted
Channel characteristics figure；
S3：The multi-channel feature figure is scanned using sliding window, preliminary screening goes out dial plate region candidate window；
S4：The relevant position feature of the dial plate region candidate window position is extracted, final target detection result is obtained；
S5：Using the loss of dial plate region candidate window and final goal loss, the parameter of the full convolution recurrent neural net is updated.
2. the water meter disc area detection method of full convolution recurrent neural network is based on as claimed in claim 1, and its feature exists
In step S1 includes step
S11：The water meter image pattern in multiple actual scenes is gathered by photographic head, the water meter image pattern includes various light
According under the conditions of, different visual angles, different type, the different extent of damages, the water meter image pattern of different rotary angle；
S12：The region of water meter disk in the water meter image pattern is labeled, including water meter disc area external world rectangle frame
Four apex region coordinate positions (x1, y1), (x2, y2), (x3, y3), (x4, y4).
3. the water meter disc area detection method of full convolution recurrent neural network is based on as claimed in claim 1, and its feature exists
In step S2 includes step
S21：Build full convolution recurrent neural network, the full convolution recurrent neural network includes multiple convolutional layers and pond layer
Cascade, is input into as triple channel RGB image, is output as multichannel characteristic pattern；
S22：By error back propagation and stochastic gradient descent method, the parameter of the full convolution recurrent neural network is carried out excellent
Change and update.
4. the water meter disc area detection method of full convolution recurrent neural network is based on as claimed in claim 1, and its feature exists
In step S3 includes step
S31：Sliding window scanning is carried out to the multi-channel feature figure, and multi-channel feature figure in sliding window is carried out into feature
Fusion；
S32：The full Connection Neural Network of multiple multilamellars is built, the full Connection Neural Network of each multilamellar is each responsible for mesh under different scale
Target is detected and positioned.
5. the water meter disc area detection method of full convolution recurrent neural network is based on as claimed in claim 1, and its feature exists
In step S4 includes step
According to the dial plate region candidate window, the feature of relevant position is extracted on the multi-channel feature figure, and carry out scale
Very little spatial pyramid pond, obtains characteristic vector, and characteristic vector obtains the aobvious of target through grader and after returning device calculating
Work property and rectangle frame parameter, the target to detecting carries out non-maximization and suppresses, and obtains detecting target.
6. the water meter disc area detection method of full convolution recurrent neural network is based on as claimed in claim 1, and its feature exists
In step S5 also includes step
S51：Water meter image in step S1 is replaced with into water meter image to be tested；
S52：Repeat step S2-S4, obtains final target detection result；
S53：Non- maximization is carried out to the final target detection result to suppress, and obtains final target detection result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611114543.1A CN106682664A (en) | 2016-12-07 | 2016-12-07 | Water meter disc area detection method based on full convolution recurrent neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611114543.1A CN106682664A (en) | 2016-12-07 | 2016-12-07 | Water meter disc area detection method based on full convolution recurrent neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106682664A true CN106682664A (en) | 2017-05-17 |
Family
ID=58868367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611114543.1A Pending CN106682664A (en) | 2016-12-07 | 2016-12-07 | Water meter disc area detection method based on full convolution recurrent neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106682664A (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107506822A (en) * | 2017-07-26 | 2017-12-22 | 天津大学 | A kind of deep neural network method based on Space integration pond |
| CN108009469A (en) * | 2017-10-24 | 2018-05-08 | 中国科学院电子学研究所苏州研究院 | A kind of offshore oil and gas detection of platform method based on structure recurrent neural network |
| CN108021914A (en) * | 2017-12-27 | 2018-05-11 | 清华大学 | Printed matter character zone extracting method based on convolutional neural networks |
| CN108280455A (en) * | 2018-01-19 | 2018-07-13 | 北京市商汤科技开发有限公司 | Human body critical point detection method and apparatus, electronic equipment, program and medium |
| CN108615010A (en) * | 2018-04-24 | 2018-10-02 | 重庆邮电大学 | Facial expression recognizing method based on the fusion of parallel convolutional neural networks characteristic pattern |
| CN108647686A (en) * | 2018-05-11 | 2018-10-12 | 同济大学 | A kind of water meter image Recognition of Reading method based on convolutional neural networks |
| CN108921203A (en) * | 2018-06-13 | 2018-11-30 | 深圳市云识科技有限公司 | A kind of detection and recognition methods of pointer-type water meter |
| CN109492580A (en) * | 2018-11-08 | 2019-03-19 | 北方工业大学 | Multi-size aerial image positioning method based on full convolution network field saliency reference |
| CN109543688A (en) * | 2018-11-14 | 2019-03-29 | 北京邮电大学 | A kind of novel meter reading detection and knowledge method for distinguishing based on multilayer convolutional neural networks |
| CN109711253A (en) * | 2018-11-19 | 2019-05-03 | 国家电网有限公司 | Ammeter technique for partitioning based on convolutional neural networks and Recognition with Recurrent Neural Network |
| CN109840497A (en) * | 2019-01-30 | 2019-06-04 | 华南理工大学 | A kind of pointer-type water meter reading detection method based on deep learning |
| CN109871754A (en) * | 2019-01-08 | 2019-06-11 | 深圳禾思众成科技有限公司 | A kind of instrument read method, equipment and computer readable storage medium |
| CN109886951A (en) * | 2019-02-22 | 2019-06-14 | 北京旷视科技有限公司 | Method for processing video frequency, device and electronic equipment |
| CN109949225A (en) * | 2019-03-11 | 2019-06-28 | 厦门美图之家科技有限公司 | A kind of image processing method and calculate equipment |
| CN110059617A (en) * | 2019-04-17 | 2019-07-26 | 北京易达图灵科技有限公司 | A kind of recognition methods of target object and device |
| CN110097600A (en) * | 2019-05-17 | 2019-08-06 | 百度在线网络技术（北京）有限公司 | The method and device of traffic mark board for identification |
| CN110516678A (en) * | 2019-08-27 | 2019-11-29 | 北京百度网讯科技有限公司 | Image processing method and device |
| CN112783327A (en) * | 2021-01-29 | 2021-05-11 | 中国科学院计算技术研究所 | Method and system for gesture recognition based on surface electromyogram signals |
| WO2021227995A1 (en) * | 2020-05-09 | 2021-11-18 | 杭州海康威视数字技术股份有限公司 | Method and apparatus for generating reference data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2648379A2 (en) * | 2012-08-17 | 2013-10-09 | Institute of Electronics, Chinese Academy of Sciences | Method and apparatus for estimating channel error |
| CN106097391A (en) * | 2016-06-13 | 2016-11-09 | 浙江工商大学 | A kind of multi-object tracking method identifying auxiliary based on deep neural network |
| CN106127204A (en) * | 2016-06-30 | 2016-11-16 | 华南理工大学 | A kind of multi-direction meter reading Region detection algorithms of full convolutional neural networks |
-
2016
- 2016-12-07 CN CN201611114543.1A patent/CN106682664A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2648379A2 (en) * | 2012-08-17 | 2013-10-09 | Institute of Electronics, Chinese Academy of Sciences | Method and apparatus for estimating channel error |
| CN106097391A (en) * | 2016-06-13 | 2016-11-09 | 浙江工商大学 | A kind of multi-object tracking method identifying auxiliary based on deep neural network |
| CN106127204A (en) * | 2016-06-30 | 2016-11-16 | 华南理工大学 | A kind of multi-direction meter reading Region detection algorithms of full convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| MING LIANG等: "Recurrent convolutional neural network for object recognition", 《2015 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)》 * |
| 宣森炎等: "基于联合卷积和递归神经网络的交通标志识别", 《传感器与微***》 * |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107506822A (en) * | 2017-07-26 | 2017-12-22 | 天津大学 | A kind of deep neural network method based on Space integration pond |
| CN107506822B (en) * | 2017-07-26 | 2021-02-19 | 天津大学 | Deep neural network method based on space fusion pooling |
| CN108009469A (en) * | 2017-10-24 | 2018-05-08 | 中国科学院电子学研究所苏州研究院 | A kind of offshore oil and gas detection of platform method based on structure recurrent neural network |
| CN108009469B (en) * | 2017-10-24 | 2020-11-24 | 中国科学院电子学研究所苏州研究院 | Offshore oil and gas platform detection method based on structure recurrent neural network |
| CN108021914B (en) * | 2017-12-27 | 2020-07-28 | 清华大学 | Method for extracting character area of printed matter based on convolutional neural network |
| CN108021914A (en) * | 2017-12-27 | 2018-05-11 | 清华大学 | Printed matter character zone extracting method based on convolutional neural networks |
| CN108280455A (en) * | 2018-01-19 | 2018-07-13 | 北京市商汤科技开发有限公司 | Human body critical point detection method and apparatus, electronic equipment, program and medium |
| CN108615010A (en) * | 2018-04-24 | 2018-10-02 | 重庆邮电大学 | Facial expression recognizing method based on the fusion of parallel convolutional neural networks characteristic pattern |
| CN108615010B (en) * | 2018-04-24 | 2022-02-11 | 重庆邮电大学 | Facial expression recognition method based on parallel convolution neural network feature map fusion |
| CN108647686A (en) * | 2018-05-11 | 2018-10-12 | 同济大学 | A kind of water meter image Recognition of Reading method based on convolutional neural networks |
| CN108921203A (en) * | 2018-06-13 | 2018-11-30 | 深圳市云识科技有限公司 | A kind of detection and recognition methods of pointer-type water meter |
| CN109492580A (en) * | 2018-11-08 | 2019-03-19 | 北方工业大学 | Multi-size aerial image positioning method based on full convolution network field saliency reference |
| CN109543688A (en) * | 2018-11-14 | 2019-03-29 | 北京邮电大学 | A kind of novel meter reading detection and knowledge method for distinguishing based on multilayer convolutional neural networks |
| CN109711253A (en) * | 2018-11-19 | 2019-05-03 | 国家电网有限公司 | Ammeter technique for partitioning based on convolutional neural networks and Recognition with Recurrent Neural Network |
| CN109871754A (en) * | 2019-01-08 | 2019-06-11 | 深圳禾思众成科技有限公司 | A kind of instrument read method, equipment and computer readable storage medium |
| CN109840497A (en) * | 2019-01-30 | 2019-06-04 | 华南理工大学 | A kind of pointer-type water meter reading detection method based on deep learning |
| CN109886951A (en) * | 2019-02-22 | 2019-06-14 | 北京旷视科技有限公司 | Method for processing video frequency, device and electronic equipment |
| CN109949225A (en) * | 2019-03-11 | 2019-06-28 | 厦门美图之家科技有限公司 | A kind of image processing method and calculate equipment |
| CN110059617A (en) * | 2019-04-17 | 2019-07-26 | 北京易达图灵科技有限公司 | A kind of recognition methods of target object and device |
| CN110097600A (en) * | 2019-05-17 | 2019-08-06 | 百度在线网络技术（北京）有限公司 | The method and device of traffic mark board for identification |
| CN110097600B (en) * | 2019-05-17 | 2021-08-06 | 百度在线网络技术（北京）有限公司 | Method and device for identifying traffic sign |
| CN110516678B (en) * | 2019-08-27 | 2022-05-06 | 北京百度网讯科技有限公司 | Image processing method and device |
| CN110516678A (en) * | 2019-08-27 | 2019-11-29 | 北京百度网讯科技有限公司 | Image processing method and device |
| US11514263B2 (en) | 2019-08-27 | 2022-11-29 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for processing image |
| WO2021227995A1 (en) * | 2020-05-09 | 2021-11-18 | 杭州海康威视数字技术股份有限公司 | Method and apparatus for generating reference data |
| CN112783327B (en) * | 2021-01-29 | 2022-08-30 | 中国科学院计算技术研究所 | Method and system for gesture recognition based on surface electromyogram signals |
| CN112783327A (en) * | 2021-01-29 | 2021-05-11 | 中国科学院计算技术研究所 | Method and system for gesture recognition based on surface electromyogram signals |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106682664A (en) | Water meter disc area detection method based on full convolution recurrent neural network | |
| CN106127204B (en) | A kind of multi-direction meter reading Region detection algorithms of full convolutional neural networks | |
| CN112766274B (en) | Water gauge image water level automatic reading method and system based on Mask RCNN algorithm | |
| CN104988928B (en) | Method for monitoring foundation pit excavation horizontal displacement in real-time manner based on digital images | |
| CN110765934B (en) | Geological disaster identification method based on multi-source data fusion | |
| CN109993094A (en) | Fault in material intelligent checking system and method based on machine vision | |
| CN109241913A (en) | In conjunction with the ship detection method and system of conspicuousness detection and deep learning | |
| CN108648169A (en) | The method and device of high voltage power transmission tower defects of insulator automatic identification | |
| CN108921203A (en) | A kind of detection and recognition methods of pointer-type water meter | |
| CN107631782B (en) | Water level detection method based on Harris angular point detection | |
| CN114972191A (en) | Method and device for detecting farmland change | |
| CN110782461A (en) | Mask-RCNN-based infrared image segmentation method for electric power equipment | |
| CN112560619B (en) | Multi-focus image fusion-based multi-distance bird accurate identification method | |
| CN109697441A (en) | A kind of object detection method, device and computer equipment | |
| CN111369539B (en) | Building facade window detecting system based on multi-feature image fusion | |
| CN109086649A (en) | Satellite remote sensing images identifying water boy method | |
| CN104121850A (en) | Canopy density measurement method and device | |
| CN107133623A (en) | A kind of pointer position accurate detecting method positioned based on background subtraction and the center of circle | |
| CN109614871A (en) | Photovoltaic roof and photovoltaic barrier automatic identification algorithm | |
| CN111861866A (en) | Panoramic reconstruction method for substation equipment inspection image | |
| CN103593858A (en) | Method for filtering out green vegetation data in ground laser radar scanning data | |
| CN109886146A (en) | Flood information remote-sensing intelligent acquisition method and equipment based on Machine Vision Detection | |
| CN109344850A (en) | A kind of water meter automatic testing method based on YOLO | |
| CN105423975A (en) | Calibration system and method of large-size workpiece | |
| CN110363706A (en) | A kind of large area bridge floor image split-joint method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170517 |
|
| WD01 | Invention patent application deemed withdrawn after publication |