CN111474863A - Weather identification model construction method, identification method and device - Google Patents
Weather identification model construction method, identification method and device Download PDFInfo
- Publication number
- CN111474863A CN111474863A CN201910064954.1A CN201910064954A CN111474863A CN 111474863 A CN111474863 A CN 111474863A CN 201910064954 A CN201910064954 A CN 201910064954A CN 111474863 A CN111474863 A CN 111474863A
- Authority
- CN
- China
- Prior art keywords
- weather
- category
- categories
- feature
- identification
- 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
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B17/00—Systems involving the use of models or simulators of said systems
- G05B17/02—Systems involving the use of models or simulators of said systems electric
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a method for constructing a weather identification model, a method for identifying the weather identification model and a device thereof, wherein the method comprises the following steps: acquiring a training set containing a plurality of weather categories, wherein the training set comprises a plurality of images corresponding to each weather category; preprocessing the weather image to obtain a feature set corresponding to each weather category; combining the feature sets corresponding to all weather categories to obtain a plurality of sample sets, wherein each sample set comprises at least two feature sets corresponding to the weather categories; and training each sample to obtain a plurality of recognition models. The embodiment of the application provides a plurality of weather identification models obtained by combining the feature sets corresponding to all weather categories, so that the weather image to be predicted can be input into the weather identification models, the quick weather identification is realized, and the speed and accuracy of the weather identification are improved.
Description
Technical Field
The application relates to the technical field of machine learning, in particular to a weather identification model construction method, a weather identification method and a weather identification device.
Background
In the logistics industry, in order to improve the timeliness of service, a large number of logistics unmanned aerial vehicles are adopted for dispatching goods. In the process of dispatching tasks executed by the logistics unmanned aerial vehicle, the weather conditions in the flight area need to be mastered in real time so as to ensure flight safety.
At present, an aviation weather identification system covers a large airplane take-off and landing airport and a airline, and during weather identification, a plurality of standard SVM models of two classifications are firstly adopted to classify weather to be identified, and finally classification results of the plurality of standard SVM models are voted to finish weather identification and prediction.
For small and medium-sized logistics unmanned aerial vehicles, weather conditions of covered air routes and flight areas are complex, when sudden or local severe weather conditions occur, a standard SVM model in the existing aviation weather identification system can only identify two weather types at each time, so that the identification process is complex, timely and effective local weather prediction cannot be provided for the unmanned aerial vehicle, and the potential safety hazard of flight exists.
Disclosure of Invention
In view of the above defects or shortcomings in the prior art, it is desirable to provide a method for constructing a weather identification model, a method for identifying a weather identification model, a device for identifying a weather identification model, and an unmanned aerial vehicle, so as to improve the speed and accuracy of weather identification.
In a first aspect, an embodiment of the present application provides a method for building a weather identification model, where the method includes:
acquiring a training set of a plurality of weather categories, wherein the training set comprises a plurality of images corresponding to the weather categories;
preprocessing the weather image to obtain a feature set corresponding to each weather category;
combining the feature sets corresponding to all weather categories to obtain a plurality of sample sets, wherein each sample set comprises at least two feature sets corresponding to the weather categories, and the number of the weather categories corresponding to the feature sets in the at least two sample sets is different;
and training each sample set to obtain the identification model corresponding to each sample set, so that each identification model separates at least one weather category from a plurality of weather categories.
In a second aspect, an embodiment of the present application provides a weather identification method, where the method includes:
acquiring an image of weather to be identified;
preprocessing the image to obtain a feature set corresponding to the weather to be identified;
inputting the feature set into the weather identification model according to the first aspect, and outputting a weather category corresponding to the weather to be identified.
In a third aspect, an embodiment of the present application provides a weather identification model building apparatus, where the apparatus includes:
the acquisition module is used for acquiring a training set of a plurality of weather categories, and the training set comprises a plurality of images corresponding to the weather categories;
the processing module is used for preprocessing the weather images to obtain a feature set corresponding to each weather category;
the combination module is used for combining the feature sets corresponding to all weather categories to obtain a plurality of sample sets, and each sample set comprises at least two feature sets corresponding to the weather categories;
and the training module is used for training the samples to obtain a plurality of recognition models, so that each recognition model separates at least one weather category from a plurality of weather categories.
In a fourth aspect, an embodiment of the present application provides a weather identification apparatus, including:
the acquisition module is used for acquiring an image of weather to be identified;
the processing module is used for preprocessing the image to obtain a feature set corresponding to the weather to be identified;
and the identification module is used for inputting the feature set into the weather identification model in the first aspect and outputting the weather category corresponding to the weather to be identified.
In summary, according to the method for constructing the weather identification model, the method for identifying the weather identification model and the device for identifying the weather identification model, the images corresponding to the multiple weather categories are obtained as the training sets, the feature sets of the images corresponding to all the weather categories are combined to obtain the multiple sample sets, so that the number of the weather categories included in at least two sample sets is different, the sample sets are trained to obtain the multiple weather identification models, the weather images to be predicted can be input into all or part of the weather identification models, the weather is quickly identified, and the speed and the accuracy of weather identification are improved.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
fig. 1 is a schematic flow chart of a method for constructing a weather identification model according to an embodiment of the present disclosure;
FIG. 2 is a schematic flow chart illustrating a method for constructing a weather identification model according to another embodiment of the present disclosure;
fig. 3 is a schematic flow chart of a method for constructing a weather identification model according to an embodiment of the present disclosure;
fig. 4 is a schematic flowchart of a weather identification method according to an embodiment of the present application;
fig. 5 is a schematic flowchart of a weather identification method according to another embodiment of the present application;
fig. 6 is a schematic structural diagram of an unmanned aerial vehicle provided in an embodiment of the present application;
fig. 7 is a schematic structural diagram of a weather identification model building apparatus according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of a weather identification apparatus according to an embodiment of the present application;
fig. 9 is a schematic structural diagram of a computer system of a server according to an embodiment of the present application.
Detailed Description
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
The method for building the weather identification model can be suitable for building the weather identification model in any scene, and the built identification model can be used for identifying the weather in any scene correspondingly.
In a logistics transportation scene, the current aviation weather identification system mainly covers a large airplane take-off and landing airport and a flight line, and the flight line and a flight area covered by a medium-sized and small-sized logistics unmanned aerial vehicle have more complex climate conditions compared with the large airplane. Therefore, when small-size commodity circulation unmanned aerial vehicle meets unexpected or local bad weather condition in the middle of, current ground aviation weather identification system can't provide timely effectual weather prediction for commodity circulation unmanned aerial vehicle, also can't provide effectual weather early warning for it yet. Under the above-mentioned scene, just need unmanned aerial vehicle self can be timely effectual discernment current weather.
The weather identification construction model, the weather identification method and the weather identification device can be used for the logistics unmanned aerial vehicle, namely the constructed weather identification model is deployed on the unmanned aerial vehicle, so that the unmanned aerial vehicle can timely and effectively identify weather in the current flight area, and safety guarantee is provided for the flight of the unmanned aerial vehicle.
For convenience of understanding and explanation, the method for constructing a weather identification model, the method for identifying weather identification model and the device provided by the embodiment of the present application are explained in detail through fig. 1 to 9.
Fig. 1 is a schematic flow chart of a method for building a weather identification model provided in an embodiment of the present application, and as shown in fig. 1, the method may include:
s110, acquiring a training set of a plurality of weather categories.
And S120, preprocessing the weather image to obtain a feature set corresponding to each weather category.
And S130, combining the feature sets corresponding to all weather categories to obtain a plurality of sample sets.
And S140, training each sample set to obtain a plurality of recognition models.
Specifically, according to the method for constructing the weather identification model provided by the embodiment of the application, images of all weather categories can be obtained as a training set. For example, a plurality of images of different weathers such as sunny weather, rain weather, snow weather, fog weather, sand and dust weather are acquired by means of a visual sensor and the like.
And then preprocessing the acquired images to obtain the characteristics corresponding to each image, namely obtaining the characteristic set corresponding to each weather category.
Alternatively, contrast features, sharpness features, texture features, hue features, saturation features, and/or brightness features of the image may be extracted.
For example, the contrast characteristic of the image can be calculated by the following formula:
wherein M and N are respectively the number of rows and columns of the image, IijIs the gray value of the ith row and the jth column pixel,
is the image mean gray value.
For the sharpness feature of an image, it can be characterized by the variance of the image gradient norm, which is shown as:
wherein G isx(m, n) and GyAnd (m and n) are respectively the convolution of the gray values of the pixel points m and n with the x-square and y-direction Sobel operators.
The variance of the image gradient norm is then:
wherein the content of the first and second substances,
is the mean of the image gradient modes.
For the texture features of the image, firstly, a co-occurrence matrix of the image gray level needs to be obtained, and the co-occurrence matrix coordinate is as follows:
wherein p, q are pixel coordinates, i, j are pixel values, and the co-occurrence matrix element represents the number of times that a particular pixel value occurs at a particular offset relationship at the same time.
It can be understood that, according to the co-occurrence matrix, the following four characteristics of the image texture can be calculated:
angular second moment:
where k is the number of gray values of the image.
Contrast of gray level co-occurrence matrix:
correlation coefficient of gray level co-occurrence matrix:
wherein, muiAnd mujIs the mean value of the gray levels; sigmaiAnd σjIs the gray scale variance.
Entropy of gray level co-occurrence matrix:
for hue characteristics, saturation characteristics and brightness characteristics, the hue characteristics, the saturation characteristics and the brightness characteristics can be represented through HSV color characteristics, namely, RGB colors are converted into HSV space, hue, saturation and brightness of each pixel are obtained, and three characteristics of image colors are represented through the hue, saturation and brightness characteristics.
By the method, at least 9 features of each image, namely 9-dimensional feature vectors, can be obtained, so that a feature set corresponding to each weather category is obtained.
And then combining the feature sets corresponding to all weather categories to obtain at least one sample set, wherein each sample set comprises at least two feature sets corresponding to the weather categories. For example, the sample set may include feature sets corresponding to two weather categories, and may further include feature sets corresponding to three or more weather categories. And, there are at least two sample sets, and the number of weather categories corresponding to the feature sets in the two sample sets is different.
For example, a binary tree model as shown in fig. 3 may be used for combination, and all weather type images included in each parent node in the binary tree shown in fig. 3 are taken as a sample set, so that each sample set includes feature sets corresponding to at least two weather categories. That is, on the basis of the sample sets including the feature sets of all weather categories, the feature sets corresponding to one or more weather categories are gradually removed, so that a plurality of sample sets can be obtained by combination, or on the basis of the feature set corresponding to one weather category, the feature sets corresponding to the weather categories are gradually added, a plurality of sample sets are obtained by combination, and the number of the weather categories included in at least two sample sets is different.
After obtaining a plurality of feature sets corresponding to different weather categories, machine training may be performed on each sample set to obtain a recognition model corresponding to each sample set, so that at least one recognition model can separate one weather category from a plurality of weather categories.
According to the method for constructing the weather identification model, the characteristic sets corresponding to all weather categories are combined to obtain a plurality of sample sets comprising different numbers of weather categories, the sample sets are trained to obtain a plurality of different identification models, and therefore when the weather is predicted by the identification models, whether the weather needs to be further input into other identification models or not can be judged according to the output result of the current identification model, and the quick identification of the weather types is achieved.
Fig. 2 is a schematic flow chart of a method for constructing a weather identification model according to another embodiment of the present application, and as shown in fig. 2, the method may include:
s210, acquiring a training set of a plurality of weather categories.
S220, preprocessing the weather image to obtain a feature set corresponding to each weather category.
It is understood that the above steps are the same as the implementation process of the embodiment shown in fig. 1, and are not described herein again.
And S231, combining the feature set corresponding to the first weather category and the feature set corresponding to the second weather category subset to serve as a first sample set.
And S232, combining the feature sets corresponding to the weather categories in the third weather sub-set and the fourth weather sub-set in the second weather sub-set to serve as a second sample set.
Specifically, the feature set corresponding to a certain weather category may be combined with the feature sets corresponding to all other weather categories to serve as a first sample set, that is, the feature sets corresponding to all weather categories serve as the first sample set.
And one of all weather categories is a first weather category, and the rest weather categories except the first weather category form a second weather category subset. That is, the corresponding feature sets in the second subset of weather categories form a second sample set, and the second subset of weather categories may include a third subset of weather categories and a fourth subset of weather categories.
For example, it is necessary to construct an identification model for three weather categories, namely, clear weather category, rain weather category and snow weather category, the first weather category may be clear weather category, and the second subset of weather categories may include two weather categories, namely rain weather category and snow weather category. Namely, the first sample set comprises feature sets corresponding to three weather categories of fine weather, rain and snow, and the second sample set only comprises feature sets corresponding to two weather categories of rain and snow.
For another example, a first sample set including feature sets of all weather categories may be obtained by combining feature sets corresponding to five different weather categories, namely, sunny weather, rain, snow, fog, and dust. And the weather category of the sunny weather can be used as a first weather category, the remaining weather categories of the rainy, snowy, foggy and sandy dust are used as a second weather category subset, namely, the feature set corresponding to the four weather categories of the rainy, snowy, foggy and sandy dust is a second sample set.
It can be understood that, in the embodiment of the present application, the execution sequence of the above-mentioned combination steps is not limited, and the combination steps may be executed synchronously or sequentially.
It is to be understood that after obtaining the first sample set and the second sample set, if the number of weather categories in the second subset of weather categories is not greater than 2, or no further identification is made on the weather categories in the second subset of weather categories, the first sample set and the second sample set may be directly trained, i.e., S241 is executed.
If the number of weather categories in the second subset of weather categories is greater than 2, in order to further identify each weather, the weather categories in the second subset of weather categories may be combined, i.e., S233 and/or S234 may be performed.
And S233, combining the feature sets corresponding to the weather categories in the third weather category subset to serve as a third sample set.
And S234, combining the feature sets corresponding to the weather categories in the fourth subset of the weather categories to obtain a fourth sample set.
Specifically, if at least two types of weather are included in the third weather category subset and/or the fourth weather category subset belonging to the second weather category subset, the feature sets in the third weather category subset and/or the fourth weather category subset are required to be used as separate training samples, that is, the third sample set and the fourth sample set.
For example, the feature set corresponding to the four middle weather categories of rain, snow, fog and sand dust is used as the second weather category subset, the weather categories of rain and snow are used as the third weather category subset, and the fog and sand dust are used as the fourth weather category subset. In order to further determine whether the weather belongs to one of the four categories of rain, snow, fog and sand dust, the feature set corresponding to the weather category of rain and snow may be used as a separate sample set, that is, a third sample set, and the feature set corresponding to the weather category of fog and sand dust may be used as a separate training sample set, that is, a fourth sample set.
For another example, if the second subset of weather categories includes three weather categories, rain, snow and fog, and the third subset of weather categories may include only one weather category, snow, and the fourth subset of weather categories includes two weather categories, rain and fog. In this case, in order to further distinguish the two weather categories of rain and fog, the feature sets corresponding to the two weather categories of rain and fog need to be used as separate training sets, that is, the fourth sample set.
And S241, training the first sample set to obtain a first recognition model.
And S242, training the second sample set to obtain a second recognition model.
Specifically, after obtaining the plurality of sample sets, each sample set may be trained to obtain a plurality of weather identification models.
It can be understood that the first sample set is trained to obtain the first recognition model, and since the first sample set includes the feature sets corresponding to the first weather category and the second weather category subset, the first recognition model may be used to indicate whether the input weather to be recognized belongs to the first category, that is, the first category weather is separated from all weather categories. Similarly, since the second sample set includes the feature sets corresponding to the third weather category subset and the fourth weather category subset, the second identification model may be used to indicate whether the input weather to be identified belongs to the weather in the third weather category subset or the fourth weather category subset.
For example, the first sample set includes a feature set corresponding to three weather categories of sunny, rainy and snowy, the first weather category is sunny, and the second weather category subset includes two weather categories of rainy and snowy, so that the obtained first identification model can identify whether the weather to be identified belongs to sunny, that is, the weather category of sunny is separated from other categories. The second sample set comprises feature sets corresponding to two weather categories, namely rain and snow, and the obtained second identification model can identify whether the weather to be identified belongs to rain or snow.
Further, if there are more weather categories in the second subset of weather categories, at least three weather categories are provided, the recognition model needs to be trained further based on the obtained sample set for the weather in the second subset of weather categories. I.e., after performing S233 and/or S234, it is necessary to perform corresponding S243 and/or S244,
and S243, training the third sample set to obtain a third recognition model.
And S244, training the fourth sample set to obtain a fourth recognition model.
Specifically, when at least two weather categories are in the third weather category subset, the obtained third sample set may be trained to obtain a third recognition model.
And when at least two weather categories exist in the fourth subset of weather categories, training the obtained fourth sample set to obtain a fourth recognition model.
It is understood that, in practice, S243 and S244 may be selectively executed according to the number of weather categories in the third weather category subset and the fourth weather category subset.
For example, the first sample set includes feature sets corresponding to four weather categories of fine, rain, snow and fog, the first weather category is fine, the second weather category subset includes three weather categories of rain, snow and fog, the third weather category subset includes two weather categories of rain and fog, and the fourth weather category subset includes only one weather category of snow. The second recognition model can recognize whether the weather to be recognized belongs to snow, i.e., the snow is separated from other weather categories. And under the scene, training a sample set of the feature set corresponding to the third weather category subset to obtain a third recognition model, wherein the third recognition model is used for accurately recognizing whether the weather category to be input belongs to rain or fog.
For another example, as shown in fig. 3, the first sample set includes feature sets corresponding to five weather categories, i.e., clear weather, rain, snow, fog, and dust, which can be respectively represented as weather 1, 2, 3, 4, and 5. And the first weather category is clear, and the second weather category subset includes four weather categories, namely rain, snow, fog and sand, so that the obtained first identification model can identify whether the weather to be identified belongs to clear, namely, the weather to be identified is separated from other weather categories on clear days. The second training set comprises feature sets corresponding to four weather types of rain, snow, fog and dust, the third weather type subset comprises two weather types of rain and snow, and the fourth weather type subset comprises two weather types of fog and dust, so that the second recognition model can recognize whether the weather to be recognized belongs to rain and snow or two weather types of fog and dust.
Under the scene, further training can be carried out on sample sets corresponding to the third weather category subset and the fourth weather category subset to obtain a third recognition model and a third recognition model so as to accurately recognize whether the weather belongs to rain and snow; fog and dust.
According to the method for constructing the weather identification model, the obtained weather types are combined in various ways, the one-to-many identification model is trained firstly to separate one weather type, then the rest weather types are divided into two types, and finally the identification models are trained to obtain the weather types to be predicted, so that the multiple identification models can be selectively utilized to sequentially identify the weather types to be predicted, and the speed and the accuracy of weather identification are improved.
It can be understood that after the identification model is obtained, the obtained identification model can be used in an actual weather identification scene.
Fig. 4 is a schematic flowchart of a weather identification method provided in an embodiment of the present application, and as shown in fig. 4, the method may include:
s410, acquiring an image of weather to be identified;
s420, preprocessing the image to obtain a feature set corresponding to the weather to be identified;
and S430, inputting the feature set into the weather identification model of the embodiment, and outputting the weather category corresponding to the weather to be identified.
Specifically, the operation method for obtaining the weather image and preprocessing the weather image in S410 and S420 to obtain the feature set corresponding to the image is the same as that in the above embodiment, and is not described here again.
The acquired feature set of the weather image to be identified is input into the identification model constructed in the embodiment, and whether the weather image needs to be further input into other identification models is judged according to the output result of the current identification model, so that the weather type is quickly identified to obtain the weather category.
In another embodiment shown in fig. 5, the method may include:
s510, acquiring an image of weather to be identified;
s520, preprocessing the image to obtain a feature set corresponding to the weather to be identified.
And S531, inputting the feature set of the weather to be identified into the first identification model, and outputting a first label of the weather to be identified.
S532, whether the first label represents that the weather to be identified belongs to the first weather category or not is judged.
And S533, inputting the feature set of the weather to be identified into the second identification model, and outputting a second label of the weather to be identified.
Specifically, the first recognition model is obtained by training a first sample set, where the first sample set includes a feature set corresponding to a first weather category and a feature set corresponding to a second weather category subset.
It is to be appreciated that the first tag can indicate whether the weather to be identified belongs to a first weather category. And if the first label indicates that the weather to be identified belongs to the first weather category, accurately obtaining the category of the weather to be identified, namely separating the first weather category from all weather categories, and ending the identification method.
Otherwise, S533 may be entered, where the feature set of the weather to be recognized needs to be further input into the second recognition model, and the second label is output. Since the second recognition model is derived from the second sample set, the second sample set includes the third weather category subset of the second weather sub-category and the feature set corresponding to the weather category of the fourth weather category subset. That is, the second label may indicate whether the weather to be identified belongs to the weather category in the third weather category subset or the weather category in the fourth weather category subset.
For example, all weather categories are clear, rain, snow. The label output by the first recognition model may indicate whether the weather to be recognized belongs to sunny, such as 1 indicating sunny, and 0 indicating non-sunny, i.e. indicating rain or snow. If the output label is 1, the method is ended, namely, the first recognition model can be used for separating sunny weather from rain and snow weather. Otherwise, the feature set of the weather image to be recognized needs to be input to the second recognition model for further recognition by the second recognition model. For example, the second tag outputs 1, indicated as rain, and 0, indicated as snow.
It is understood that if only one weather category is included in the third subset of weather categories and the fourth subset of weather categories, the identification method ends, otherwise S535 and/or S536 need to be performed.
Further, if the second label can indicate that the weather to be identified belongs to the weather category in the third subset of weather categories, S535 is performed, and if the second label can indicate that the weather to be identified belongs to the weather category in the fourth subset of weather categories, S536 is performed.
S534, judging whether the second label represents that the weather to be identified belongs to the weather category in the third weather category subset or the weather category in the fourth weather category subset.
And S535, inputting the feature set of the weather to be recognized into a third recognition model, and outputting a third label of the weather to be recognized.
And S536, inputting the feature set of the weather to be identified into a fourth identification model, and outputting a fourth label of the weather to be identified.
For example, in the scene shown in fig. 3, all weather categories are weather of sunny, rainy, snowy, fog, and dust. Namely, the first sample set comprises a feature set corresponding to five weather categories of fine weather, rain, snow, fog and sand dust, the first weather category is fine, the second weather category subset comprises four weather categories of rain, snow, fog and sand dust, the third weather category subset comprises two weather categories of rain and snow, and the fourth weather category subset comprises two weather categories of fog and sand dust.
When the second label output by the second recognition model indicates that the weather to be recognized belongs to rain or snow, the feature set of the weather image to be recognized needs to be input into the third recognition model, that is, S435 is performed, so that the third recognition model outputs the final category of the weather to be recognized.
When the second tag output by the second recognition model indicates that the weather to be recognized belongs to fog or dust, the feature set of the weather image to be recognized needs to be input into the fourth recognition model, that is, S436 is executed, so that the third recognition model outputs the final category of the weather to be recognized.
For another example, all weather categories are three types of weather, such as sunny, rainy, snowy, and dusty. And the first weather category is clear, the second weather category subset comprises three weather types of rain, snow and sand, the third weather category subset comprises two weather types of rain and snow, and the fourth weather category subset comprises a weather category of sand and dust.
S535 needs to be performed when the second tag output by the second recognition model indicates that the weather to be recognized belongs to the third weather category.
When the second label output by the second recognition model indicates that the weather to be recognized belongs to the fourth category of weather, the method ends.
The embodiment of the application provides a weather identification method, which can input a first identification model at first through a pre-constructed weather identification model, and determine whether to input a next identification model according to an output result of the identification model, so that the complexity of weather identification can be reduced, and the weather identification speed can be increased.
The weather identification model building and identification method can be used in the logistics industry, and when the unmanned aerial vehicle carries out cargo transportation, the weather in the flight area of the unmanned aerial vehicle is identified, so that real-time effective weather early warning is provided for the navigation of the unmanned aerial vehicle.
As shown in fig. 6, an embodiment of the present application further provides an unmanned aerial vehicle 600, where the unmanned aerial vehicle includes a body 610, and further includes a weather identification system 620, and the weather identification system may be configured to perform the weather identification method described above.
Preferably, in the unmanned aerial vehicle provided by the embodiment of the present application, the weather identification system may further perform the method for constructing the weather identification model.
Preferably, the unmanned aerial vehicle provided in the embodiment of the present application further includes a vision sensor 630, configured to acquire an image of weather to be identified and/or acquire a training set of at least one weather category.
Preferably, the unmanned aerial vehicle provided by the embodiment of the application further comprises an early warning device 640, which is used for generating an early warning signal based on the tag of the weather to be identified.
Preferably, unmanned aerial vehicle that this application embodiment provided still includes:
and the control device 650 is used for generating a control instruction based on the early warning signal, and the control instruction is used for controlling the unmanned aerial vehicle to land or return.
Specifically, after the weather identification system identifies the weather condition of the current flight area, the weather identification system can send the weather condition to the ground and perform ground analysis. Or the identification tag based on the current weather in the unmanned aerial vehicle can be analyzed, and if the analysis result shows that the flight safety cannot be met, the warning device 640 of the aircraft can feed back a prompt message, so that the warning device 640 generates a warning signal and sends the warning signal to the control device. The control device can generate a control instruction according to the early warning signal. The control command can start the landing assembly to enable the unmanned aerial vehicle to land safely; or, the control instruction changes the unmanned aerial vehicle route so that the unmanned aerial vehicle safely navigates back.
The unmanned aerial vehicle that this application embodiment provided is through the image that obtains current flight area to in the weather identification model of deployment on unmanned aerial vehicle with the characteristic set input of this image, make unmanned aerial vehicle at the flight in-process, realize weather identification fast, improved unmanned aerial vehicle to the speed and the rate of accuracy of weather identification under the current area, ensured unmanned aerial vehicle's flight safety
Fig. 7 is a schematic structural diagram of a weather identification model building apparatus according to an embodiment of the present application, and as shown in fig. 7, the apparatus 700 may include:
an obtaining module 710, configured to obtain a training set of at least one weather category, where the training set includes at least one image corresponding to the weather category;
the processing module 720 is configured to pre-process the weather image to obtain a feature set corresponding to each weather category;
the combination module 730 is configured to combine the feature sets corresponding to all the weather categories to obtain at least one sample set, where the sample set includes feature sets corresponding to at least two weather categories, and the number of the weather categories corresponding to the feature sets in the at least two sample sets is different;
the training module 740 is configured to train the at least one sample to obtain at least one recognition model, so that the at least one recognition model separates a weather category from a plurality of weather categories.
Preferably, in the weather identification model building apparatus provided in the embodiment of the present application, the combining module 730 includes:
a first combining unit 731, configured to combine a feature set corresponding to the first weather category and a feature set corresponding to a second weather category subset as a first sample set, where the second weather category subset includes at least two weather categories;
a second combining unit 732, configured to combine feature sets corresponding to weather categories in the third weather sub-set and the fourth weather sub-set in the second weather sub-category to obtain a second sample set, where the third weather sub-set and the fourth weather sub-set at least include one weather category.
Preferably, in the weather identification model building apparatus provided in the embodiment of the present application, the combination model 730 further includes:
a third combining unit 733, configured to combine feature sets corresponding to weather categories in the third weather category subset to obtain a third sample set;
and/or
A fourth combining unit 734, configured to combine the feature sets corresponding to the weather categories in the fourth subset of weather categories to obtain a fourth sample set; wherein the content of the first and second substances,
the third weather category subset and the fourth weather category subset include at least two weather categories.
Preferably, in the apparatus for constructing a weather recognition model provided in the embodiment of the present application, the training module 740 includes:
a first training unit 741, configured to train the first sample set to obtain a first recognition model, where the first recognition model is used to indicate whether the weather to be predicted belongs to a first category.
A second training unit 742, configured to train the second sample set to obtain a second recognition model, where the second recognition model is used to indicate that the weather to be predicted belongs to a weather category in the third weather category subset or the fourth weather category subset.
Preferably, in the weather identification model building apparatus provided in the embodiment of the present application, the training module 740 further includes:
a third training unit 743, configured to train the third sample set to obtain a third recognition model, where the third recognition model is used to indicate that the weather to be predicted belongs to one of the weather categories in the third weather category subset.
A fourth training unit 744, configured to train the fourth sample set to obtain a fourth recognition model, where the fourth recognition model is used to indicate that the weather to be predicted belongs to one of the weather categories in the fourth subset of weather categories.
Preferably, in the weather identification model building apparatus provided in the embodiment of the present application, the processing module 720 is specifically configured to:
and extracting the contrast characteristic, the sharpness characteristic, the texture characteristic, the hue characteristic, the saturation characteristic and/or the brightness characteristic of the weather image to obtain a characteristic set corresponding to each weather category.
Fig. 7 is a schematic structural diagram of a weather identification apparatus according to an embodiment of the present application, and as shown in fig. 7, the apparatus 800 may include:
and an obtaining module 810, configured to obtain an image of weather to be identified.
And the processing module 820 is configured to pre-process the image to obtain a feature set corresponding to the weather to be identified.
The identifying module 830 is configured to input the feature set into the weather identification model as described above, and output a weather category corresponding to the weather to be identified.
Preferably, in the apparatus for constructing a weather identification model provided in the embodiment of the present application, the identification module 830 includes:
the first identification unit 831 is used for inputting the feature set of the weather to be identified into the first identification model and outputting a first label of the weather to be identified; the first identification model is obtained by training a first sample set, and the first sample set comprises a feature set corresponding to a first weather category and a feature set corresponding to a second weather category subset;
a determining unit 832, configured to determine whether the first tag indicates that the weather to be identified belongs to a first weather category;
the second identifying unit 833 is configured to, when the first tag indicates that the weather to be identified does not belong to the first weather category, input the feature set of the weather to be identified into a second identifying model, and output a second tag of the weather to be identified, where the second identifying model is obtained from a second sample set, and the second sample set includes feature sets corresponding to weather categories in a third weather sub-category and a fourth weather sub-category in the second weather sub-category.
Preferably, in the weather identification model building apparatus provided in the embodiment of the present application, the identification module 830 further includes:
a third identifying unit 834, configured to input the feature set of the weather to be identified into a third identifying model, and output a third tag of the weather to be identified, where the third identifying model is obtained from a third sample set, and the third sample set includes a feature set corresponding to a weather category in the third weather category subset;
alternatively, the first and second electrodes may be,
the fourth identifying unit 835 inputs the feature set of the weather to be identified into a fourth identifying model, and outputs a fourth tag of the weather to be identified, where the fourth identifying model is obtained from a fourth sample set, and the fourth sample set includes the feature set corresponding to the weather category in the fourth subset of the weather category.
It is understood that the embodiments of the present application also provide a server, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the program to implement the weather identification model building and the weather identification method as described above.
Referring now to FIG. 9, shown is a block diagram of a computer system 900 suitable for use in implementing a server according to embodiments of the present application.
As shown in fig. 9, the computer system 900 includes a Central Processing Unit (CPU)901 that can execute various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)902 or a program loaded from a storage section 909 into a Random Access Memory (RAM) 909. In the RAM 909, various programs and data necessary for the operation of the system 900 are also stored. The CPU 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input/output (I/O) interface 905 is also connected to bus 904.
To the I/O interface 905, AN input section 906 including a keyboard, a mouse, and the like, AN output section 907 including a device such as a Cathode Ray Tube (CRT), a liquid crystal display (L CD), and the like, a speaker, and the like, a storage section 908 including a hard disk, and the like, and a communication section 909 including a network interface card such as a L AN card, a modem, and the like, the communication section 909 performs communication processing via a network such as the internet, a drive 910 is also connected to the I/O interface 905 as necessary, a removable medium 911 such as a magnetic disk, AN optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 910 as necessary, so that a computer program read out therefrom is mounted into the storage section 908 as necessary.
In particular, the processes described above with reference to fig. 1-5 may be implemented as computer software programs, according to embodiments disclosed herein. For example, embodiments disclosed herein include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising program code for performing the method of FIG. 1. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 909, and/or installed from the removable medium 911.
It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various weather identification embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units or modules described in the embodiments of the present application may be implemented by software or hardware. The described units or modules may also be provided in a processor, and may be described as: a processor includes an acquisition module, a processing module, a combination module, and a training module. The names of the units or modules do not in some cases form a limitation on the units or modules themselves, for example, the combination module may also be described as a "module for combining feature sets corresponding to all weather categories to obtain at least one sample set".
As another aspect, the present application also provides a computer-readable storage medium, which may be the computer-readable storage medium included in the foregoing device in the foregoing embodiment; or it may be a separate computer readable storage medium not incorporated into the device. The computer readable storage medium stores one or more programs for use by one or more processors in performing the weather identification model construction described herein, in particular: acquiring a training set of at least one weather category, wherein the training set comprises at least one image corresponding to the weather category; preprocessing the weather image to obtain a feature set corresponding to each weather category; combining the feature sets corresponding to all weather categories to obtain at least one sample set, wherein the sample set comprises at least two feature sets of the weather categories, and the number of the weather categories corresponding to the feature sets in the at least two sample sets is different; and training the at least one sample to obtain at least one recognition model.
To sum up, the weather identification model construction method, the weather identification method and the weather identification device provided by the embodiment of the application combine the feature sets of the images corresponding to all weather categories by acquiring the images corresponding to the weather categories as a training set to obtain a plurality of sample sets, so that the number of the weather categories included in at least two sample sets is different, the sample sets are trained to obtain a plurality of weather identification models, the weather images to be predicted can be input into all or part of the weather identification models deployed on the unmanned aerial vehicle, the unmanned aerial vehicle can quickly realize weather identification in the flight process, the speed and the accuracy of the unmanned aerial vehicle for weather identification in the current area are improved, and the flight safety of the unmanned aerial vehicle is ensured.
The above description is only a preferred embodiment of the application and is illustrative of the principles of the technology employed. It will be appreciated by a person skilled in the art that the scope of the invention as referred to in the present application is not limited to the embodiments with a specific combination of the above-mentioned features, but also covers other embodiments with any combination of the above-mentioned features or their equivalents without departing from the inventive concept. For example, the above features may be replaced with (but not limited to) features having similar functions disclosed in the present application.
Claims (12)
1. A method for constructing a weather identification model is characterized by comprising the following steps:
acquiring a training set of a plurality of weather categories, wherein the training set comprises a plurality of images corresponding to the weather categories;
preprocessing a weather image to obtain a feature set corresponding to each weather category;
combining feature sets corresponding to all weather categories to obtain a plurality of sample sets, wherein the sample sets comprise feature sets of at least two weather categories, and the number of the weather categories corresponding to the feature sets in at least two sample sets is different;
training each sample set to obtain a recognition model corresponding to the sample set, so that each recognition model separates at least one weather category from a plurality of weather categories.
2. The weather identification model building method of claim 1, wherein the combining the feature sets corresponding to all weather categories to obtain a plurality of sample sets comprises:
combining a feature set corresponding to a first weather category and a feature set corresponding to a second weather category subset as a first sample set, wherein the second weather category subset comprises at least two weather categories;
and combining the feature sets corresponding to the weather categories in the third weather sub-set and the fourth weather sub-set in the second weather sub-set to serve as a second sample set, wherein the third weather sub-set and the fourth weather sub-set at least comprise one weather category.
3. The method of building a weather prediction model of claim 2, further comprising:
combining feature sets corresponding to the weather categories in the third weather category subset to serve as a third sample set;
and/or
Combining feature sets corresponding to the weather categories in the fourth subset of weather categories to serve as a fourth sample set; wherein the content of the first and second substances,
and the third weather category subset and the fourth weather category subset at least comprise two weather categories.
4. The method of claim 2, wherein the training of each sample set to obtain the recognition model corresponding to the sample set comprises:
training the first sample set to obtain a first identification model, wherein the first identification model is used for indicating whether the weather to be predicted belongs to a first category or not;
and training the second sample set to obtain a second identification model, wherein the second identification model is used for indicating that the weather to be predicted belongs to a weather category in a third weather category subset or a fourth weather category subset.
5. The method of claim 3, wherein the training of each sample set to obtain the recognition model corresponding to the sample set comprises:
training the third sample set to obtain a third recognition model, wherein the third recognition model is used for indicating that the weather to be predicted belongs to one of weather categories in a third weather category subset;
and training the fourth sample set to obtain a fourth recognition model, wherein the fourth recognition model is used for indicating that the weather to be predicted belongs to one of weather categories in a fourth subset of weather categories.
6. The method of claim 1, wherein the preprocessing the weather image to obtain the feature set corresponding to each weather category comprises:
and extracting contrast characteristics, sharpness characteristics, texture characteristics, hue characteristics, saturation characteristics and/or brightness characteristics of the weather image to obtain a characteristic set corresponding to each weather category.
7. A weather identification method, the method comprising:
acquiring an image of weather to be identified;
preprocessing the image to obtain a feature set corresponding to the weather to be identified;
inputting the feature set into the weather identification model according to any one of claims 1 to 6, and outputting a weather category corresponding to the weather to be identified.
8. The weather identification method according to claim 7, wherein the inputting the feature set into the identification model according to any one of claims 1 to 6, and the outputting the weather category corresponding to the weather to be identified comprises:
inputting a feature set of weather to be identified into a first identification model, and outputting a first label of the weather to be identified; the first identification model is obtained by training a first sample set, and the first sample set comprises a feature set corresponding to a first weather category and a feature set corresponding to a second weather category subset;
judging whether the first label represents that the weather to be identified belongs to a first weather category or not;
if not, the feature set of the weather to be identified is input into a second identification model, and a second label of the weather to be identified is output, wherein the second identification model is obtained from a second sample set, and the second sample set comprises a third weather category subset in the second weather subcategory and a feature set corresponding to a weather category in a fourth weather category subset.
9. The weather identification method of claim 8, further comprising:
inputting the feature set of the weather to be identified into a third identification model, and outputting a third label of the weather to be identified, wherein the third identification model is obtained from a third sample set, and the third sample set comprises a feature set corresponding to a weather category in the third weather category subset;
alternatively, the first and second electrodes may be,
inputting the feature set of the weather to be identified into a fourth identification model, wherein the fourth identification model is obtained from a fourth sample set, and the fourth sample set comprises a feature set corresponding to a weather category in the fourth subset of the weather category.
10. A weather recognition model construction apparatus, characterized in that the apparatus comprises:
the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring a training set of a plurality of weather categories, and the training set comprises a plurality of images corresponding to the weather categories;
the processing module is used for preprocessing the weather image to obtain a feature set corresponding to each weather category;
the combination module is used for combining the feature sets corresponding to all weather categories to obtain a plurality of sample sets, wherein the sample sets comprise feature sets corresponding to at least two weather categories, and the number of the weather categories corresponding to the feature sets in the at least two sample sets is different;
and the training module is used for training each sample set to obtain the identification model corresponding to the sample set, so that each identification model separates at least one weather category from a plurality of weather categories.
11. The weather recognition model building apparatus of claim 10, wherein the combining module comprises:
the first combination unit is used for combining a feature set corresponding to a first weather category and a feature set corresponding to a second weather category subset as a first sample set, wherein the second weather category subset comprises at least two weather categories;
and the second combination unit is used for combining the feature sets corresponding to the weather categories in the third weather sub-set and the fourth weather sub-set in the second weather sub-category to serve as a second sample set, and the third weather sub-set and the fourth weather sub-set at least comprise one weather category.
12. A weather identification device, the device comprising:
the acquisition module is used for acquiring an image of weather to be identified;
the processing module is used for preprocessing the image to obtain a feature set corresponding to the weather to be identified;
the identification module is used for inputting the feature set into the weather identification model as claimed in any one of claims 1 to 6 and outputting the weather category corresponding to the weather to be identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910064954.1A CN111474863A (en) | 2019-01-23 | 2019-01-23 | Weather identification model construction method, identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910064954.1A CN111474863A (en) | 2019-01-23 | 2019-01-23 | Weather identification model construction method, identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111474863A true CN111474863A (en) | 2020-07-31 |
Family
ID=71743316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910064954.1A Pending CN111474863A (en) | 2019-01-23 | 2019-01-23 | Weather identification model construction method, identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111474863A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463196A (en) * | 2014-11-11 | 2015-03-25 | 中国人民解放军理工大学 | Video-based weather phenomenon recognition method |
| CN105868745A (en) * | 2016-06-20 | 2016-08-17 | 重庆大学 | Weather identifying method based on dynamic scene perception |
| CN108875593A (en) * | 2018-05-28 | 2018-11-23 | 上海交通大学 | Visible images weather recognition methods based on convolutional neural networks |
| CN108897073A (en) * | 2018-06-20 | 2018-11-27 | 西安电子科技大学 | Weather prediction method based on Beidou signal and terrestrial wireless signal |
-
2019
- 2019-01-23 CN CN201910064954.1A patent/CN111474863A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463196A (en) * | 2014-11-11 | 2015-03-25 | 中国人民解放军理工大学 | Video-based weather phenomenon recognition method |
| CN105868745A (en) * | 2016-06-20 | 2016-08-17 | 重庆大学 | Weather identifying method based on dynamic scene perception |
| CN108875593A (en) * | 2018-05-28 | 2018-11-23 | 上海交通大学 | Visible images weather recognition methods based on convolutional neural networks |
| CN108897073A (en) * | 2018-06-20 | 2018-11-27 | 西安电子科技大学 | Weather prediction method based on Beidou signal and terrestrial wireless signal |
Non-Patent Citations (7)
| Title |
|---|
| GJORGJI MADZAROV, DEJAN GJORGJEVIKJ: "Multi-class classification using support vector machines in decision tree architecture", no. 10791143, pages 288 - 295, XP031491863 * |
| YUGUO WANG: "A Tree-based Multi-class SVM Classifier for Digital Library Document", vol. 10720540, pages 15 - 18, XP031476925 * |
| 侯慧蓉: "基于特征选择的天气现象识别方法研究", 中国优秀硕士学位论文全文数据库信息科技辑, vol. 2018, no. 02, pages 138 - 2534 * |
| 张天坤: "基于机器视觉的天气现象识别研究与实现", 中国优秀硕士学位论文全文数据库信息科技辑, vol. 2015, no. 04, pages 138 - 746 * |
| 张天坤: "基于机器视觉的天气识别研究与实现", vol. 2015, no. 4, pages 138 - 746 * |
| 李骞，范茵，张璟，李宝强: "基于室外图像的天气现象识别方法", vol. 31, no. 06, pages 1624 - 1627 * |
| 王建芬,曹元大: "支持向量机在大类别数分类中的应用", vol. 2001, no. 02, pages 225 - 228 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11783568B2 (en) | Object classification using extra-regional context | |
| WO2019223582A1 (en) | Target detection method and system | |
| EP4091109A1 (en) | Systems for multiclass object detection and alerting and methods therefor | |
| US11651302B2 (en) | Method and device for generating synthetic training data for an artificial-intelligence machine for assisting with landing an aircraft | |
| CN112712057B (en) | Traffic signal identification method and device, electronic equipment and storage medium | |
| US20210158157A1 (en) | Artificial neural network learning method and device for aircraft landing assistance | |
| CN102073846A (en) | Method for acquiring traffic information based on aerial images | |
| CN110969171A (en) | Image classification model, method and application based on improved convolutional neural network | |
| EP4089645A1 (en) | Aircraft classification from aerial imagery | |
| CN111461006B (en) | Optical remote sensing image tower position detection method based on deep migration learning | |
| Nagarani et al. | Unmanned Aerial vehicle’s runway landing system with efficient target detection by using morphological fusion for military surveillance system | |
| CN112668441A (en) | Satellite remote sensing image airplane target identification method combined with priori knowledge | |
| CN115049948A (en) | Unmanned aerial vehicle inspection method and device based on neural network model and related equipment | |
| CN113011528B (en) | Remote sensing image small target detection method based on context and cascade structure | |
| CN114048536A (en) | Road structure prediction and target detection method based on multitask neural network | |
| CN117152706A (en) | Aircraft runway accumulated water identification method, device and system | |
| CN111474863A (en) | Weather identification model construction method, identification method and device | |
| CN115035331A (en) | Small sample target detection method and system applied to small unmanned aerial vehicle-mounted platform | |
| JP2011081614A (en) | Recognition system, recognition method, and program | |
| CN115909245A (en) | Visual multi-task processing method based on deep learning | |
| Dhulipudi et al. | Geospatial object detection using machine learning-aviation case study | |
| Skuratov et al. | Application of kohonen self-organizing map to search for region of interest in the detection of objects | |
| Volkov et al. | Devising an image processing method for transport infrastructure monitoring systems | |
| CN110555437B (en) | Image semantic segmentation model, detection method, detection device, detection equipment and storage medium thereof | |
| Tsekhmystro et al. | Study of methods for searching and localizing objects in images from aircraft using convolutional neural networks |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210719 Address after: 518063 5th floor, block B, building 1, software industry base, Yuehai street, Nanshan District, Shenzhen City, Guangdong Province Applicant after: Fengyi Technology (Shenzhen) Co.,Ltd. Address before: 518061 Intersection of Xuefu Road (south) and Baishi Road (east) in Nanshan District, Shenzhen City, Guangdong Province, 6-13 floors, Block B, Shenzhen Software Industry Base Applicant before: SF TECHNOLOGY Co.,Ltd. |
|
| TA01 | Transfer of patent application right |