CN108932510A - A kind of rubbish detection method and device - Google Patents

A kind of rubbish detection method and device Download PDF

Info

Publication number
CN108932510A
CN108932510A CN201810951193.7A CN201810951193A CN108932510A CN 108932510 A CN108932510 A CN 108932510A CN 201810951193 A CN201810951193 A CN 201810951193A CN 108932510 A CN108932510 A CN 108932510A
Authority
CN
China
Prior art keywords
rubbish
image
detected
detection method
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
Application number
CN201810951193.7A
Other languages
Chinese (zh)
Inventor
廖海斌
肖春红
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guizhou Is Suitable For Zhitong Technology Co Ltd
Original Assignee
Guizhou Is Suitable For Zhitong Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guizhou Is Suitable For Zhitong Technology Co Ltd filed Critical Guizhou Is Suitable For Zhitong Technology Co Ltd
Priority to CN201810951193.7A priority Critical patent/CN108932510A/en
Publication of CN108932510A publication Critical patent/CN108932510A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/10Terrestrial scenes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/40Extraction of image or video features

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Image Analysis (AREA)

Abstract

The present invention provides a kind of rubbish detection method and device, are related to image procossing and technical field of environmental detection.The rubbish detection method acquires the image to be detected of first area in situation to be detected first, feature extraction is carried out to benchmark image of the first area in no rubbish and obtains first eigenvector, feature extraction is carried out to described image to be detected and obtains second feature vector, determine that the similarity of the first eigenvector and the second feature vector is greater than default similar threshold value, then judged in described image to be detected using convolutional neural networks object detection model with the presence or absence of rubbish image, if so, determining that there are rubbish for the first area.The rubbish detection method carries out automatic detection, identification and the alarm of rubbish by image procossing and neural network model, improves the efficiency and accuracy of rubbish detection, reduces cost of labor.

Description

A kind of rubbish detection method and device
Technical field
The present invention relates to image procossing and technical field of environmental detection, in particular to a kind of rubbish detection method and Device.
Background technique
With the quickening of ur- banization speed and the sustainable growth of population size, the scale of cell is gradually expanded, and cell enters Firmly rate also improves rapidly, and population density's increase then brings huge challenge to cell environment.It safeguards that cell is clean and tidy, keeps Graceful environment becomes particularly important.Cell environment, which becomes the comprehensive of cell overall image, to be embodied, and the life of resident is directly affected Quality and health living.And cell rubbish be influence cell environment a key element, cell road surface rubbish monitoring with Tidiness evaluation becomes the important content of cell regulatory authorities routine work.
Cell pavement garbage monitoring and evaluation is manually patrolled and registration of taking pictures mainly by special messenger at present.This artificial detection Method needs to expend a large amount of manpower and material resources, and is had a holiday by weather, personnel and influenced with the various extraneous factors such as working time, There is a problem of that low efficiency, there may be mistake registration, human cost are high.
Summary of the invention
In view of this, the embodiment of the present invention is designed to provide a kind of rubbish detection method and device, it is above-mentioned to solve Low efficiency present in existing artificial rubbish detection method, there may be the high problems of mistake registration, human cost.
In a first aspect, the rubbish detection method acquires first the embodiment of the invention provides a kind of rubbish detection method The image to be detected of first area in situation to be detected carries out benchmark image of the first area in no rubbish Feature extraction obtains first eigenvector, carries out feature extraction to described image to be detected and obtains second feature vector, determines institute The similarity for stating first eigenvector and the second feature vector is less than default similar threshold value, using convolutional neural networks object Detection model judges with the presence or absence of rubbish image in described image to be detected, if so, determining that there are rubbish for the first area.
It is comprehensive in a first aspect, in the case that it is described to the first area in no rubbish benchmark image carry out feature extraction Before obtaining first eigenvector, the rubbish detection method further include: acquire the first area in no rubbish Benchmark image.
It is comprehensive in a first aspect, in the acquisition first area after image to be detected in situation to be detected, Yi Jisuo State benchmark image to the first area in no rubbish and carry out feature extraction and obtain first eigenvector, to it is described to Before detection image carries out feature extraction acquisition second feature vector, the rubbish detection method further include: to the reference map Picture and described image to be detected carry out size normalization processing.
It is comprehensive in a first aspect, it is described judged using convolutional neural networks object detection model be in described image to be detected Before the no image there are rubbish, the rubbish detection method further include: acquire several described first areas respectively in no rubbish feelings Under condition and there is the image in the case of rubbish as training sample;Training set is obtained after carrying out data extending to the training sample; Convolutional neural networks object detection model training is carried out based on the training set, obtains convolutional neural networks object detection model.
Synthesis is in a first aspect, obtain training set after the progress data extending to the training sample, comprising: to the instruction Practice sample to carry out mirror image, invert, cut out, obtaining training set after the data extending of color disturbance.
Synthesis is in a first aspect, described carry out convolutional neural networks object detection model training based on the training set, comprising: The pre- instruction of garbage classification task is carried out to the archetype of convolutional neural networks object detection model on ImageNet data set Practice, the initial model of the convolutional neural networks object detection model is obtained according to the model parameter value that the pre-training obtains; Arameter optimization is carried out to the initial model using the training set, obtains convolutional neural networks object detection model.
Synthesis is in a first aspect, the rubbish detection method is also wrapped in the determination first area there are after rubbish It includes: determining the grade of rubbish alarm according to position, type and the size of rubbish in the rubbish image;To the sending pair of default object Answer the rubbish alarm of grade.
Synthesis is in a first aspect, determine that rubbish is alert in position, type and the size according to rubbish in the rubbish image Before the grade of report, the rubbish detection method further include: described image to be detected is divided into multiple subregions, each subregion point Different rubbish position weight value l is not corresponded to, identifies the subregion of rubbish corresponding position in described image to be detected, is obtained The corresponding rubbish position weight value l of the subregion;Calculate the ratio s of the rubbish image Yu image to be detected size;It will Different rubbish types is correspondingly arranged different rubbish class label c, is identified in the rubbish image using depth network model Rubbish type;The rubbish class parameter θ that the rubbish is obtained by rubbish alert level formula θ=c (l+s), according to described The affiliated threshold interval of rubbish class parameter θ determines the grade of rubbish alarm.
Second aspect, the embodiment of the invention provides a kind of rubbish detection device, the rubbish detection device includes image Obtain module, characteristic extracting module, similarity judgment module and rubbish judgment module.Described image obtains module for obtaining the The image to be detected of one region in situation to be detected.The characteristic extracting module is used for the first area in no rubbish feelings Benchmark image under condition carries out feature extraction and obtains first eigenvector, carries out feature extraction to described image to be detected and obtains the Two feature vectors.The similarity judgment module is used to determine the similar of the first eigenvector and the second feature vector Degree is greater than default similar threshold value.The rubbish judgment module be used for using the judgement of convolutional neural networks object detection model it is described to It whether there is rubbish image in detection image, if so, further determining that rubbish classification.
Comprehensive second aspect, the rubbish detection device further include that training sample obtains module, training set obtains module and Convolutional neural networks object detection model training module.The training sample obtain module for obtain respectively several described first Region is in no rubbish and has the image in the case of rubbish as training sample.The training set obtains module and is used for institute It states after training sample carries out data extending and obtains training set.The convolutional neural networks object detection model training module is used for base Convolutional neural networks object detection model training is carried out in the training set, obtains convolutional neural networks object detection model.
Comprehensive second aspect, the rubbish detection device further include position weight value computing module, image size calculating mould Block, rubbish category identification module and rubbish alert level determining module.The position weight value computing module be used for will it is described to Detection image is divided into multiple subregions, and each subregion respectively corresponds different rubbish position weight value l, identifies the rubbish in institute The subregion for stating corresponding position in image to be detected obtains the corresponding rubbish position weight value l of the subregion.The big subtotal of described image Calculate the ratio s that module is used to calculate the rubbish image Yu image to be detected size.The rubbish category identification module is used In different rubbish types to be correspondingly arranged to different rubbish class label c, the rubbish image is identified using depth network model In rubbish type.The rubbish alert level determining module is used to obtain by rubbish alert level formula θ=c (l+s) The rubbish class parameter θ of the rubbish determines the grade of rubbish alarm according to the affiliated threshold interval of the rubbish class parameter θ.
The third aspect, the embodiment of the invention also provides a kind of storage medium, the storage medium is stored in computer, The storage medium includes a plurality of instruction, and a plurality of instruction is configured such that the computer executes above-mentioned method.
Beneficial effect provided by the invention is:
The present invention provides a kind of rubbish detection method, the rubbish detection method is using first area in no rubbish situation Under image carry out image procossing as image to be detected for needing to carry out rubbish detection of benchmark image and subsequent acquisition, be based on Image procossing and neural network model carry out rubbish detection by computer automatically, improve the efficiency of rubbish detection and accurate Property, reduce cost of labor.Wherein, the rubbish detection method using neural network model carry out the judgement of rubbish image before, Also the feature vector similarity of the benchmark image and described image to be detected is compared, it is pre- to determine that the similarity is less than If just executing the rubbish image judgement when threshold value, avoid directlying adopt neural network model progress rubbish image judgement and wasting A large amount of computing resource further improves the efficiency and resource utilization of the rubbish detection.Meanwhile using convolutional Neural net Network object detection model is judged with the presence or absence of rubbish image in described image to be detected, can pass through convolutional neural networks Frame in object detection model, which is returned, carries out identification and frame choosing to the rubbish image in described image to be detected, improves rubbish The accuracy of detection.
Other features and advantages of the present invention will be illustrated in subsequent specification, also, partly be become from specification It is clear that by implementing understanding of the embodiment of the present invention.The objectives and other advantages of the invention can be by written theory Specifically noted structure is achieved and obtained in bright book, claims and attached drawing.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this A little attached drawings obtain other relevant attached drawings.
Fig. 1 is a kind of flow diagram for rubbish detection method that first embodiment of the invention provides;
Fig. 2 is a kind of process for convolutional neural networks object detection model training step that first embodiment of the invention provides Schematic diagram;
Fig. 3 is a kind of flow diagram for rubbish alarm step that first embodiment of the invention provides;
Fig. 4 is a kind of module diagram for rubbish detection device that second embodiment of the invention provides;
Fig. 5 is a kind of structure that can be applied to the electronic equipment in the embodiment of the present application that third embodiment of the invention provides Block diagram.
Icon: 100- rubbish detection device;110- image collection module;120- characteristic extracting module;130- similarity is sentenced Disconnected module;140- rubbish judgment module;200- electronic equipment;201- memory;202- storage control;203- processor; 204- Peripheral Interface;205- input-output unit;206- audio unit;207- display unit.
Specific embodiment
Below in conjunction with attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete Ground description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Usually exist The component of the embodiment of the present invention described and illustrated in attached drawing can be arranged and be designed with a variety of different configurations herein.Cause This, is not intended to limit claimed invention to the detailed description of the embodiment of the present invention provided in the accompanying drawings below Range, but it is merely representative of selected embodiment of the invention.Based on the embodiment of the present invention, those skilled in the art are not doing Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing.Meanwhile of the invention In description, term " first ", " second " etc. are only used for distinguishing description, are not understood to indicate or imply relative importance.
First embodiment
Through the applicant the study found that existing rubbish detection method relies primarily on artificial progress, manually patrols and take pictures Registration.This artificial detection method needs to expend a large amount of manpower and material resources, and had a holiday with the working time by weather, personnel etc. The influence of various extraneous factors.With domestic and international each big city " smart city ", the especially development of " intelligence community " work, such as What improves the informationization of cell environment monitoring and evaluation, intelligent level becomes an important problem.To solve the above-mentioned problems, First embodiment of the invention provides a kind of rubbish detection method, with improve rubbish detection the degree of automation, detection efficiency and Accuracy.
Referring to FIG. 1, Fig. 1 is a kind of flow diagram for rubbish detection method that first embodiment of the invention provides.
Step S10: it is to be detected in situation to be detected that detection service device controls image acquisition device first area Image.
Step S20: the detection service device carries out feature to benchmark image of the first area in no rubbish It extracts and obtains first eigenvector, feature extraction is carried out to described image to be detected and obtains second feature vector.
Step S30: the detection service device determines the similarity of the first eigenvector and the second feature vector Less than default similar threshold value.
Step S40: the detection service device judges described image to be detected using convolutional neural networks object detection model In whether there is rubbish image.
Step S50: if so, determining that there are rubbish for the first area.
For step S10, the first area can be designated area, specified community or other specified regions.The inspection Survey server can be computer, image processing workstations, Cloud Server or other have image procossing and basic operation function The processing equipment of energy.Described image acquisition device can be cell monitoring camera or unmanned plane of taking photo by plane etc. can to obtain region complete The photographic device of whole image, it is noted that the image of described image acquisition device acquisition all should be the first area Image without dead angle, all standing.Wherein, the situation to be detected does not carry out specially rubbish cleaning as, needs to carry out rubbish inspection The case where survey.
For step S20, the benchmark image is rubbish cleaning specially to be carried out to the first area and by artificial inspection It is obtained in the case that confirmation is without rubbish after looking by cell monitoring camera or unmanned plane of taking photo by plane.Optionally, the feature mentions Take step that can carry out by Gabor filter.
For step S30, the similar preset threshold can be according to the benchmark image, the resolution of described image to be detected Rate and the image-capable of the detection service device are adjusted.If point of the benchmark image, described image to be detected Resolution is higher or the image-capable of the detection service device is higher, then can be by the similar preset threshold adaptability tune It is high.
For step S40, the rubbish figure in described image to be detected is identified using convolutional neural networks object detection model Picture, and the rubbish frames images are elected, determined in identification and when frame selects rubbish image the first area there are rubbish, Determine that there is no rubbish for the first area when failing identification and frame selects rubbish image.
The present embodiment realizes the automation of the detection of the rubbish in cell or fixed area by step S10-S50, reduces The human cost of artificial carry out rubbish detection, and the benchmark image is first based on before carrying out specific rubbish image recognition Described image to be detected is tentatively judged with the presence or absence of rubbish with the similarity of the feature vector of described image to be detected, into And the waste of unnecessary computing resource is avoided, improve the efficiency of rubbish detection;Further, above-mentioned steps also pass through volume Product neural network object detection model carries out frame choosing and identification to rubbish image, improves the accuracy rate of rubbish identification.
It should be understood that as an implementation, carrying out step S10 and step S20 to the benchmark in first time Before image is handled, first area described in detection service device control image acquisition device is further comprised the steps of: in no rubbish Benchmark image in the case of rubbish.Wherein, the resolution ratio of the benchmark image, color depth, distortion situation etc. should with it is described to be checked Altimetric image is consistent as far as possible.
As an implementation, before executing step S20 and carrying out characteristic vector pickup, the present embodiment is also to the base Quasi- image and described image to be detected are pre-processed, to improve the accuracy of subsequent similarity-rough set, the pretreated step It suddenly can be with are as follows: the detection service device carries out size normalization processing to the benchmark image and described image to be detected.Wherein, In view of the resolution requirements of speed and rubbish the detection image processing of normalized, normalization size can be set to 256* 256。
Optionally, for step S30, it may be assumed that determine the similarity of the first eigenvector and the second feature vector Less than default similar threshold value, by taking benchmark image A and image to be detected B as an example, benchmark image A and image to be detected B are being carried out greatly Small normalization and feature extraction obtain the feature vector I of benchmark image AtrAnd the feature vector I of image to be detected Bte, utilize Cosine similarity calculates ItrWith IteBetween similarity:If the similarity obtained is similar less than presetting Threshold value then determines in described image to be detected there are rubbish, on the contrary then assert that there is no rubbish in described image to be detected.
It should be understood that the present embodiment judges institute using convolutional neural networks object detection model in execution step S40 It states in image to be detected with the presence or absence of before rubbish image, training is needed to obtain the convolutional neural networks object detection model. As an implementation, referring to FIG. 2, Fig. 2 is a kind of convolutional neural networks object inspection that first embodiment of the invention provides The flow diagram of model training step is surveyed, the convolutional neural networks object detection model training step can be such that
Step S31: the detection service device control described image acquisition device acquires several described first areas respectively and exists In the case of no rubbish and there is the image in the case of rubbish as training sample.
Step S32: the detection service device obtains training set after carrying out data extending to the training sample.
Step S33: the detection service device is based on the training set and carries out convolutional neural networks object detection model training, Obtain convolutional neural networks object detection model.
For step S32, optionally, the mode of the data extending can for mirror image, invert, cut out, color disturbance etc., The data extending scale of the present embodiment is 10 times, and data extending scale can be adjusted according to specific needs in other embodiments It is whole.It works to avoid carrying out Image Acquisition in large quantities by the step S32, is reducing Image Acquisition workload, improving rubbish Model training precision is improved while rubbish detection efficiency.
Optionally, for step S33, described " the detection service device is based on the training set and carries out convolutional neural networks Object detection model training obtains convolutional neural networks object detection model " may include following sub-step: the detection service Device carries out the pre- of garbage classification task to the archetype of convolutional neural networks object detection model on ImageNet data set Training, the introductory die of the convolutional neural networks object detection model is obtained according to the model parameter value that the pre-training obtains Type;The detection service device carries out arameter optimization to the initial model using the training set, obtains convolutional neural networks object Body detection model.Obtain convolutional neural networks object detection model using ImageNet data set and training set, save into Oneself carries out the time of image acquisition and mark when row initial model generates, and greatly improves convolutional neural networks object detection The training effectiveness of model, while improving the adaptability and accuracy of convolutional neural networks object detection model.
There are rubbish images in image to be detected in view of confirming the first area, i.e., there are rubbish for the described first area When rubbish, need to send corresponding notice or alarm to related personnel according to the concrete condition of rubbish.As an implementation, originally Embodiment further includes rubbish alarm step after step S40.Referring to FIG. 3, Fig. 3 is one that first embodiment of the invention provides The flow diagram of kind rubbish alarm step, step can be such that
Step S50: the detection service device determines rubbish according to position, type and the size of rubbish in the rubbish image The grade of alarm.
Step S60: the detection service device issues the rubbish alarm of corresponding grade to default object.
Optionally, for step S50, specific steps be can be such that
Step S51: described image to be detected is divided into multiple subregions by the detection service device, and each subregion respectively corresponds Different rubbish position weight value l identifies the subregion of rubbish corresponding position in described image to be detected, obtains described point The corresponding rubbish position weight value l in area.
Step S52: the detection service device calculates the ratio s of the rubbish image Yu image to be detected size.
Step S53: different rubbish types is correspondingly arranged different rubbish class label c by the detection service device, is used Depth network model identifies the rubbish type in the rubbish image.
Step S54: the detection service device obtains the rubbish of the rubbish by rubbish alert level formula θ=c (l+s) Rubbish class parameter θ determines the grade of rubbish alarm according to the affiliated threshold interval of the rubbish class parameter θ.
For step S51, multiple subregions described in the present embodiment can be public area, angle when being divided for cell It falls and garbage pool, the rubbish position weight value l being corresponding in turn to is respectively 0.6,0.3 and 0.1.It should be understood that above-mentioned subregion Division can be adjusted according to the specific requirements of cell or community, corresponding rubbish position weight value l can also change therewith Become.
Optionally, the depth network model can be GoogleNet depth network model, and before step S53, It also needs to be trained the depth network model, training step can be with are as follows: the detection service device controls described image Acquisition device acquires the rubbish image of a large amount of first areas, and marks out its rubbish type;The detection service device is to institute State rubbish image carry out mirror image, overturn, cut out, color disturbance etc. completes data extending to obtaining training set;The detection clothes Business device carries out the pre-training of classification task on ImageNet data set to the archetype of depth network model, according to described pre- The model parameter value that training obtains obtains the initial model of the depth network model;The detection service device uses the training Collection carries out arameter optimization to the initial model, obtains depth network model.
Optionally, c=4 when the rubbish type in step S53 is Harmful Waste, rubbish type are garbage or can be recycled C=2 when rubbish, c=1 when rubbish type is Other Waste.
Further, described in operator can voluntarily adjust according to the concrete type of the first area and specific requirements The specific value of rubbish class label and the rubbish position weight value.
Optionally, for step S54, the grade of rubbish alarm is determined according to rubbish class parameter θ, works as θ > θ1When be determined as Level-one alarm, which sends out a warning and presets object to level-one, sends primary alarm (serious), works as θ2≤θ≤θ1When be determined as that secondary alarm is bright Amber light simultaneously presets object transmission second-level alarm (medium) to second level, as θ < θ2When be determined as three-level alarm brilliant blue lamp and pre- to three-level If object sends three-level alarm (slight), wherein θ1, θ2For the threshold values set according to concrete condition, setting rule is θ12, together Shi Suoshu level-one presets the default object of object, second level, the default object of three-level can be set as property contact person according to specific requirements, Cleaning group contact person etc..
The present embodiment completes the judgement to rubbish type by above embodiment, reduces the detection of later period rubbish and rubbish The workload of classification, and determine according to the type of the rubbish, position and size the extent of injury of the rubbish, and to correlation Personnel send rubbish alarm corresponding with the extent of injury, thus to detecting that there are can when rubbish for the first area Timely information is carried out to corresponding personnel to feed back, to improve garbage disposal response speed, improves the environment of first area Clean and tidy degree.
Second embodiment
In order to cooperate the rubbish detection method in first embodiment of the invention, second embodiment of the invention additionally provides one kind Rubbish detection device 100.
Referring to FIG. 4, Fig. 4 is a kind of module diagram for rubbish detection device that second embodiment of the invention provides.
Rubbish detection device 100 includes image collection module 110, characteristic extracting module 120, similarity judgment module 130 With rubbish judgment module 140.
Image collection module 110 is for obtaining the image to be detected of first area in situation to be detected.Feature extraction mould Block 120 is used for the benchmark image to the first area in no rubbish and carries out feature extraction acquisition first eigenvector, Feature extraction is carried out to described image to be detected and obtains second feature vector.Similarity judgment module 130 is for determining described the One feature vector and the similarity of the second feature vector are less than default similar threshold value.Rubbish judgment module 140 is for using Convolutional neural networks object detection model judges in described image to be detected with the presence or absence of rubbish image, if so, determining described the There are rubbish in one region.
As an implementation, rubbish detection device 100 further includes that training sample obtains module, training set obtains module With convolutional neural networks object detection model training module.The training sample obtain module for obtain respectively several described the One region is in no rubbish and has the image in the case of rubbish as training sample.The training set obtain module for pair The training sample obtains training set after carrying out data extending.The convolutional neural networks object detection model training module is used for Convolutional neural networks object detection model training is carried out based on the training set, obtains convolutional neural networks object detection model.
Further, in order to carry out classification and the automatic alarm of rubbish, rubbish detection device 100 can also include that position is weighed Weight values computing module, image size computing module, rubbish category identification module and rubbish alert level determining module.The position Weight value calculation module is used to described image to be detected being divided into multiple subregions, and each subregion respectively corresponds different rubbish positions Weighted value l is set, identifies the subregion of rubbish corresponding position in described image to be detected, obtains the corresponding rubbish of the subregion Position weight value l.Described image size computing module is used to calculate the ratio of the rubbish image and image to be detected size Value s.The rubbish category identification module, for different rubbish types to be correspondingly arranged to different rubbish class label c, using depth Degree network model identifies the rubbish type in the rubbish image.The rubbish alert level determining module is used for alert by rubbish Hierarchy equation θ=c (l+s) is reported to obtain the rubbish class parameter θ of the rubbish, according to the affiliated threshold of the rubbish class parameter θ Value section determines the grade of rubbish alarm.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description Specific work process, no longer can excessively be repeated herein with reference to the corresponding process in preceding method.
3rd embodiment
Referring to figure 5., Fig. 5 is a kind of electronics that can be applied in the embodiment of the present application that third embodiment of the invention provides The structural block diagram of equipment.Electronic equipment 200 may include rubbish detection device 100, memory 201, storage control 202, place Manage device 203, Peripheral Interface 204, input-output unit 205, audio unit 206, display unit 207.
The memory 201, storage control 202, processor 203, Peripheral Interface 204, input-output unit 205, sound Frequency unit 206, each element of display unit 207 are directly or indirectly electrically connected between each other, to realize the transmission or friendship of data Mutually.It is electrically connected for example, these elements can be realized between each other by one or more communication bus or signal wire.The rubbish Detection device 100 include at least one can be stored in the form of software or firmware (firmware) in the memory 201 or The software function module being solidificated in the operating system (operating system, OS) of rubbish detection device 100.The processing Device 203 is for executing the executable module stored in memory 201, such as the software function mould that rubbish detection device 100 includes Block or computer program.
Wherein, memory 201 may be, but not limited to, random access memory (Random Access Memory, RAM), read-only memory (Read Only Memory, ROM), programmable read only memory (Programmable Read-Only Memory, PROM), erasable read-only memory (Erasable Programmable Read-Only Memory, EPROM), Electricallyerasable ROM (EEROM) (Electric Erasable Programmable Read-Only Memory, EEPROM) etc.. Wherein, memory 201 is for storing program, and the processor 203 executes described program after receiving and executing instruction, aforementioned Method performed by the server that the stream process that any embodiment of the embodiment of the present invention discloses defines can be applied to processor 203 In, or realized by processor 203.
Processor 203 can be a kind of IC chip, the processing capacity with signal.Above-mentioned processor 203 can To be general processor, including central processing unit (Central Processing Unit, abbreviation CPU), network processing unit (Network Processor, abbreviation NP) etc.;Can also be digital signal processor (DSP), specific integrated circuit (ASIC), Ready-made programmable gate array (FPGA) either other programmable logic device, discrete gate or transistor logic, discrete hard Part component.It may be implemented or execute disclosed each method, step and the logic diagram in the embodiment of the present invention.General processor It can be microprocessor or the processor 203 be also possible to any conventional processor etc..
Various input/output devices are couple processor 203 and memory 201 by the Peripheral Interface 204.Some In embodiment, Peripheral Interface 204, processor 203 and storage control 202 can be realized in one single chip.Other one In a little examples, they can be realized by independent chip respectively.
Input-output unit 205 realizes user and the server (or local terminal) for being supplied to user input data Interaction.The input-output unit 205 may be, but not limited to, the equipment such as mouse and keyboard.
Audio unit 206 provides a user audio interface, may include one or more microphones, one or more raises Sound device and voicefrequency circuit.
Display unit 207 provides an interactive interface (such as user's operation circle between the electronic equipment 200 and user Face) or for display image data give user reference.In the present embodiment, the display unit 207 can be liquid crystal display Or touch control display.It can be the capacitance type touch control screen or resistance of support single-point and multi-point touch operation if touch control display Formula touch screen etc..Single-point and multi-point touch operation is supported to refer to that touch control display can sense on the touch control display one Or at multiple positions simultaneously generate touch control operation, and the touch control operation that this is sensed transfer to processor 203 carry out calculate and Processing.
It is appreciated that structure shown in fig. 5 is only to illustrate, the electronic equipment 200 may also include more than shown in Fig. 5 Perhaps less component or with the configuration different from shown in Fig. 5.Each component shown in Fig. 5 can use hardware, software Or combinations thereof realize.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description Specific work process, no longer can excessively be repeated herein with reference to the corresponding process in preceding method.
In conclusion the rubbish detection method uses first the embodiment of the invention provides a kind of rubbish detection method The image to be detected for needing to carry out rubbish detection of image of the region in no rubbish as benchmark image and subsequent acquisition Image procossing is carried out, rubbish detection is carried out by computer based on image procossing and neural network model automatically, improves rubbish The efficiency and accuracy of detection, reduce cost of labor.Wherein, the rubbish detection method is carried out using neural network model Before the judgement of rubbish image, also the feature vector similarity of the benchmark image and described image to be detected is compared, is determined The similarity just executes the rubbish image judgement when being greater than preset threshold, avoid directlying adopt neural network model progress rubbish Rubbish image judges and wastes a large amount of computing resource, further improves the efficiency and resource utilization of the rubbish detection.Together When, judged with the presence or absence of rubbish image using convolutional neural networks object detection model in described image to be detected, energy Enough the rubbish image in described image to be detected is known by the frame recurrence in convolutional neural networks object detection model It is not selected with frame, improves the accuracy of rubbish detection.
In several embodiments provided herein, it should be understood that disclosed device and method can also pass through Other modes are realized.The apparatus embodiments described above are merely exemplary, for example, flow chart and block diagram in attached drawing Show the device of multiple embodiments according to the present invention, the architectural framework in the cards of method and computer program product, Function and operation.In this regard, each box in flowchart or block diagram can represent the one of a module, section or code Part, a part of the module, section or code, which includes that one or more is for implementing the specified logical function, to be held Row instruction.It should also be noted that function marked in the box can also be to be different from some implementations as replacement The sequence marked in attached drawing occurs.For example, two continuous boxes can actually be basically executed in parallel, they are sometimes It can execute in the opposite order, this depends on the function involved.It is also noted that every in block diagram and or flow chart The combination of box in a box and block diagram and or flow chart can use the dedicated base for executing defined function or movement It realizes, or can realize using a combination of dedicated hardware and computer instructions in the system of hardware.
In addition, each functional module in each embodiment of the present invention can integrate one independent portion of formation together Point, it is also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention. And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-OnlyMemory), arbitrary access are deposited The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.It should also be noted that similar label and letter exist Similar terms are indicated in following attached drawing, therefore, once being defined in a certain Xiang Yi attached drawing, are then not required in subsequent attached drawing It is further defined and explained.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain Lid is within protection scope of the present invention.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that There is also other identical elements in process, method, article or equipment including the element.

Claims (10)

1. a kind of rubbish detection method, which is characterized in that the rubbish detection method includes:
Acquire the image to be detected of first area in situation to be detected;
Feature extraction is carried out to benchmark image of the first area in no rubbish and obtains first eigenvector, to described Image to be detected carries out feature extraction and obtains second feature vector;
Determine that the similarity of the first eigenvector and the second feature vector is less than default similar threshold value;
Judged using convolutional neural networks object detection model with the presence or absence of rubbish image in described image to be detected, if so, really There are rubbish for the fixed first area.
2. rubbish detection method according to claim 1, which is characterized in that it is described to the first area in no rubbish In the case of benchmark image carry out feature extraction obtain first eigenvector before, the rubbish detection method further include:
Acquire benchmark image of the first area in no rubbish.
3. rubbish detection method according to claim 1, which is characterized in that in the acquisition first area in feelings to be detected After image to be detected under condition and the benchmark image to the first area in no rubbish carries out feature and mentions Acquisition first eigenvector is taken, before carrying out feature extraction acquisition second feature vector to described image to be detected, the rubbish Detection method further include:
Size normalization processing is carried out to the benchmark image and described image to be detected.
4. rubbish detection method according to claim 1, which is characterized in that examined described using convolutional neural networks object Before survey model judges to whether there is rubbish image in described image to be detected, the rubbish detection method further include:
Several described first areas are acquired respectively in no rubbish and have the image in the case of rubbish as training sample;
Training set is obtained after carrying out data extending to the training sample;
Convolutional neural networks object detection model training is carried out based on the training set, obtains convolutional neural networks object detection mould Type.
5. rubbish detection method according to claim 4, which is characterized in that described to carry out data expansion to the training sample Training set is obtained after filling, comprising:
Mirror image is carried out to the training sample, inverts, cut out, obtain training set after the data extending of color disturbance.
6. rubbish detection method according to claim 4, which is characterized in that described to carry out convolution mind based on the training set Through network object detection model training, comprising:
Garbage classification task is carried out to the archetype of convolutional neural networks object detection model on ImageNet data set Pre-training obtains the initial model of convolutional neural networks object detection model according to the model parameter value that the pre-training obtains;
Arameter optimization is carried out to the initial model using the training set, obtains convolutional neural networks object detection model.
7. rubbish detection method according to claim 1, which is characterized in that in the determination first area, there are rubbish After rubbish, the rubbish detection method further include:
The grade of rubbish alarm is determined according to position, type and the size of rubbish in the rubbish image;
The rubbish alarm of corresponding grade is issued to default object.
8. rubbish detection method according to claim 7, which is characterized in that described according to rubbish in the rubbish image Position, type and size determine rubbish alarm grade before, the rubbish detection method further include:
Described image to be detected is divided into multiple subregions, each subregion respectively corresponds different rubbish position weight value l, identification The subregion of rubbish corresponding position in described image to be detected obtains the corresponding rubbish position weight value l of the subregion;
Calculate the ratio s of the rubbish image Yu image to be detected size;
Different rubbish types is correspondingly arranged to different rubbish class label c, the rubbish figure is identified using depth network model Rubbish type as in;
The rubbish class parameter θ that the rubbish is obtained by rubbish alert level formula θ=c (l+s), according to described rubbish etc. The grade affiliated threshold interval of parameter θ determines the grade of rubbish alarm.
9. a kind of rubbish detection device, which is characterized in that the rubbish detection device includes:
Image collection module obtains the image to be detected of first area in situation to be detected;
Characteristic extracting module obtains for carrying out feature extraction to benchmark image of the first area in no rubbish One feature vector carries out feature extraction to described image to be detected and obtains second feature vector;
Similarity judgment module, for determining that it is default that the similarity of the first eigenvector and the second feature vector is less than Similar threshold value;
Rubbish judgment module, for judging to whether there is in described image to be detected using convolutional neural networks object detection model Rubbish image, if so, determining that there are rubbish for the first area.
10. a kind of storage medium, which is characterized in that be stored with computer program instructions, the computer in the storage medium When program instruction is read and run by a processor, perform claim requires the step in any one of 1-8 the method.
CN201810951193.7A 2018-08-20 2018-08-20 A kind of rubbish detection method and device Pending CN108932510A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810951193.7A CN108932510A (en) 2018-08-20 2018-08-20 A kind of rubbish detection method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810951193.7A CN108932510A (en) 2018-08-20 2018-08-20 A kind of rubbish detection method and device

Publications (1)

Publication Number Publication Date
CN108932510A true CN108932510A (en) 2018-12-04

Family

ID=64446178

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810951193.7A Pending CN108932510A (en) 2018-08-20 2018-08-20 A kind of rubbish detection method and device

Country Status (1)

Country Link
CN (1) CN108932510A (en)

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109948506A (en) * 2019-03-14 2019-06-28 南通大学 A kind of multi-angle garbage classification cloud platform based on deep learning
CN110108765A (en) * 2019-05-28 2019-08-09 浙江科技学院 Bagged garbage dry and wet recognition methods
CN110427838A (en) * 2019-07-13 2019-11-08 恒大智慧科技有限公司 A kind of rubbish method for cleaning, computer equipment and readable storage medium storing program for executing
CN110490142A (en) * 2019-08-21 2019-11-22 北京百度网讯科技有限公司 For generating the method and device of alarm signal
CN110659622A (en) * 2019-09-27 2020-01-07 北京文安智能技术股份有限公司 Detection method, device and system for garbage dumping
CN110688945A (en) * 2019-09-26 2020-01-14 成都睿云物联科技有限公司 Cleanliness detection method and device, computer equipment and storage medium
CN110956104A (en) * 2019-11-20 2020-04-03 河南华衍智能科技有限公司 Method, device and system for detecting overflow of garbage can
CN111017429A (en) * 2019-11-20 2020-04-17 重庆特斯联智慧科技股份有限公司 Community garbage classification method and system based on multi-factor fusion
CN111151368A (en) * 2020-01-09 2020-05-15 珠海格力电器股份有限公司 Garbage treatment method, system, storage medium and garbage treatment equipment
CN111767822A (en) * 2020-06-23 2020-10-13 浙江大华技术股份有限公司 Garbage detection method and related equipment and device
CN112183460A (en) * 2020-10-20 2021-01-05 武汉光谷联合集团有限公司 Method and device for intelligently identifying environmental sanitation
CN112298844A (en) * 2019-07-29 2021-02-02 杭州海康威视数字技术股份有限公司 Garbage classification monitoring method and device
CN112315383A (en) * 2020-10-29 2021-02-05 上海高仙自动化科技发展有限公司 Inspection cleaning method and device for robot, robot and storage medium
CN112668423A (en) * 2020-12-18 2021-04-16 平安科技(深圳)有限公司 Corridor sundry detection method and device, terminal equipment and storage medium
CN112686162A (en) * 2020-12-31 2021-04-20 北京每日优鲜电子商务有限公司 Method, device, equipment and storage medium for detecting clean state of warehouse environment
JP2021159881A (en) * 2020-04-01 2021-10-11 Jx金属株式会社 Composition analysis method for electronic/electrical equipment component scrap, disposal method for electronic/electrical equipment component scrap, analysis device for electronic/electrical equipment component scrap, and processing device for electronic/electrical equipment component scrap
CN114821264A (en) * 2022-03-28 2022-07-29 慧之安信息技术股份有限公司 Algorithm efficiency improving method based on neural network
WO2022174541A1 (en) * 2021-02-20 2022-08-25 北京市商汤科技开发有限公司 Garbage detection method and apparatus, device, storage medium, and program product
JP2023510063A (en) * 2020-12-18 2023-03-13 平安科技(深▲せん▼)有限公司 Method, apparatus, terminal device and storage medium for detecting debris in corridor

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105787506A (en) * 2016-01-26 2016-07-20 耿春茂 Method for assessing garbage classification based on image identification and two dimensional identification technology
CN106203498A (en) * 2016-07-07 2016-12-07 中国科学院深圳先进技术研究院 A kind of City scenarios rubbish detection method and system
CN106358021A (en) * 2016-11-01 2017-01-25 成都宏软科技实业有限公司 Smart community monitoring system
CN106845408A (en) * 2017-01-21 2017-06-13 浙江联运知慧科技有限公司 A kind of street refuse recognition methods under complex environment
CN107092914A (en) * 2017-03-23 2017-08-25 广东数相智能科技有限公司 Refuse classification method, device and system based on image recognition
CN107145842A (en) * 2017-04-19 2017-09-08 西安电子科技大学 With reference to LBP characteristic patterns and the face identification method of convolutional neural networks
CN107866386A (en) * 2017-09-30 2018-04-03 四川绿能环保科技股份公司 Perishable rubbish identifying system and method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105787506A (en) * 2016-01-26 2016-07-20 耿春茂 Method for assessing garbage classification based on image identification and two dimensional identification technology
CN106203498A (en) * 2016-07-07 2016-12-07 中国科学院深圳先进技术研究院 A kind of City scenarios rubbish detection method and system
CN106358021A (en) * 2016-11-01 2017-01-25 成都宏软科技实业有限公司 Smart community monitoring system
CN106845408A (en) * 2017-01-21 2017-06-13 浙江联运知慧科技有限公司 A kind of street refuse recognition methods under complex environment
CN107092914A (en) * 2017-03-23 2017-08-25 广东数相智能科技有限公司 Refuse classification method, device and system based on image recognition
CN107145842A (en) * 2017-04-19 2017-09-08 西安电子科技大学 With reference to LBP characteristic patterns and the face identification method of convolutional neural networks
CN107866386A (en) * 2017-09-30 2018-04-03 四川绿能环保科技股份公司 Perishable rubbish identifying system and method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
MOHAMMAD SAEED RAD: "A Computer Vision System to Localize and Classify Wastes on the Streets", 《INTERNATIONAL CONFERENCE ON COMPUTER VISION SYSTEM》 *
魏书法: "基于图像的城市场景垃圾自动检测", 《集成技术》 *

Cited By (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109948506A (en) * 2019-03-14 2019-06-28 南通大学 A kind of multi-angle garbage classification cloud platform based on deep learning
CN110108765A (en) * 2019-05-28 2019-08-09 浙江科技学院 Bagged garbage dry and wet recognition methods
CN110108765B (en) * 2019-05-28 2021-08-03 浙江科技学院 Bagged garbage dry and wet identification method
CN110427838A (en) * 2019-07-13 2019-11-08 恒大智慧科技有限公司 A kind of rubbish method for cleaning, computer equipment and readable storage medium storing program for executing
CN112298844B (en) * 2019-07-29 2023-09-22 杭州海康威视数字技术股份有限公司 Garbage classification supervision method and device
CN112298844A (en) * 2019-07-29 2021-02-02 杭州海康威视数字技术股份有限公司 Garbage classification monitoring method and device
CN110490142A (en) * 2019-08-21 2019-11-22 北京百度网讯科技有限公司 For generating the method and device of alarm signal
CN110490142B (en) * 2019-08-21 2022-08-23 北京百度网讯科技有限公司 Method and device for generating alarm signal
CN110688945A (en) * 2019-09-26 2020-01-14 成都睿云物联科技有限公司 Cleanliness detection method and device, computer equipment and storage medium
CN110659622A (en) * 2019-09-27 2020-01-07 北京文安智能技术股份有限公司 Detection method, device and system for garbage dumping
CN110956104A (en) * 2019-11-20 2020-04-03 河南华衍智能科技有限公司 Method, device and system for detecting overflow of garbage can
CN111017429A (en) * 2019-11-20 2020-04-17 重庆特斯联智慧科技股份有限公司 Community garbage classification method and system based on multi-factor fusion
CN111151368A (en) * 2020-01-09 2020-05-15 珠海格力电器股份有限公司 Garbage treatment method, system, storage medium and garbage treatment equipment
JP7301783B2 (en) 2020-04-01 2023-07-03 Jx金属株式会社 Composition analysis method for electronic/electrical equipment parts scrap, electronic/electrical equipment parts scrap processing method, electronic/electrical equipment parts scrap composition analysis device, and electronic/electrical equipment parts scrap processing equipment
JP2021159881A (en) * 2020-04-01 2021-10-11 Jx金属株式会社 Composition analysis method for electronic/electrical equipment component scrap, disposal method for electronic/electrical equipment component scrap, analysis device for electronic/electrical equipment component scrap, and processing device for electronic/electrical equipment component scrap
CN111767822B (en) * 2020-06-23 2023-04-25 浙江大华技术股份有限公司 Garbage detection method, related equipment and device
CN111767822A (en) * 2020-06-23 2020-10-13 浙江大华技术股份有限公司 Garbage detection method and related equipment and device
CN112183460A (en) * 2020-10-20 2021-01-05 武汉光谷联合集团有限公司 Method and device for intelligently identifying environmental sanitation
CN112315383A (en) * 2020-10-29 2021-02-05 上海高仙自动化科技发展有限公司 Inspection cleaning method and device for robot, robot and storage medium
CN112668423B (en) * 2020-12-18 2024-05-28 平安科技(深圳)有限公司 Corridor sundry detection method and device, terminal equipment and storage medium
WO2022126908A1 (en) * 2020-12-18 2022-06-23 平安科技(深圳)有限公司 Corridor sundries detection method and apparatus, and terminal device and storage medium
CN112668423A (en) * 2020-12-18 2021-04-16 平安科技(深圳)有限公司 Corridor sundry detection method and device, terminal equipment and storage medium
JP2023510063A (en) * 2020-12-18 2023-03-13 平安科技(深▲せん▼)有限公司 Method, apparatus, terminal device and storage medium for detecting debris in corridor
JP7258188B2 (en) 2020-12-18 2023-04-14 平安科技(深▲せん▼)有限公司 Method, apparatus, terminal device and storage medium for detecting debris in corridor
CN112686162A (en) * 2020-12-31 2021-04-20 北京每日优鲜电子商务有限公司 Method, device, equipment and storage medium for detecting clean state of warehouse environment
CN112686162B (en) * 2020-12-31 2023-12-15 鄂尔多斯市空港大数据运营有限公司 Method, device, equipment and storage medium for detecting clean state of warehouse environment
WO2022174541A1 (en) * 2021-02-20 2022-08-25 北京市商汤科技开发有限公司 Garbage detection method and apparatus, device, storage medium, and program product
CN114821264B (en) * 2022-03-28 2023-02-17 慧之安信息技术股份有限公司 Algorithm efficiency improving method based on neural network
CN114821264A (en) * 2022-03-28 2022-07-29 慧之安信息技术股份有限公司 Algorithm efficiency improving method based on neural network

Similar Documents

Publication Publication Date Title
CN108932510A (en) A kind of rubbish detection method and device
CN111126252B (en) Swing behavior detection method and related device
Taubenböck et al. Object-based feature extraction using high spatial resolution satellite data of urban areas
CN109871730A (en) A kind of target identification method, device and monitoring device
CN104134080B (en) A kind of road foundation collapses automatic testing method and system with slope failure
CN109344894A (en) Garbage classification recognition methods and device based on Multi-sensor Fusion and deep learning
CN106681996B (en) The method and apparatus for determining interest region in geographic range, point of interest
CN109201514A (en) Waste sorting recycle method, garbage classification device and classified-refuse recovery system
CN105975929A (en) Fast pedestrian detection method based on aggregated channel features
CN102902960B (en) Leave-behind object detection method based on Gaussian modelling and target contour
CN108460362A (en) A kind of system and method for detection human body
CN113822247B (en) Method and system for identifying illegal building based on aerial image
CN102880873A (en) Personnel behavior identification implementation system and method based on image segmentation and semantic extraction
KR101937940B1 (en) Method of deciding cpted cctv position by big data
WO2022174541A1 (en) Garbage detection method and apparatus, device, storage medium, and program product
CN110348437A (en) It is a kind of based on Weakly supervised study with block the object detection method of perception
CN105334224B (en) Automate quality testing cloud platform
CN102867183A (en) Method and device for detecting littered objects of vehicle and intelligent traffic monitoring system
CN111210399A (en) Imaging quality evaluation method, device and equipment
CN110490150A (en) A kind of automatic auditing system of picture violating the regulations and method based on vehicle retrieval
CN105323024A (en) Network signal intensity detecting and fusing method
CN103065118A (en) Image blurring detection method and device
CN104361357A (en) Photo set classification system and method based on picture content analysis
CN107764233A (en) A kind of measuring method and device
CN110084232A (en) The recognition methods of chinese character, device and terminal device in license plate

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
AD01 Patent right deemed abandoned
AD01 Patent right deemed abandoned

Effective date of abandoning: 20210409