Data processing method, device, electronic device and storage medium

Abstract

The present disclosure provides a data processing method, device, electronic device and storage medium, which relate to the field of data processing technology, and in particular to the field of intelligent transportation technology. The specific implementation scheme is: obtaining annotation data for a traffic light image; if the occlusion attribute in the annotation data indicates that the traffic light is occluded, determining the correction method corresponding to the occlusion attribute according to the target description information in the annotation data; wherein the target description information is used to characterize the occlusion condition of the traffic light; and correcting the occlusion attribute according to the determined correction method. It can be seen that the present disclosure uses a correction method determined by the target description information to correct the occlusion attribute. Compared with the manual correction method, it can improve the efficiency of correcting the occlusion attribute in the annotation data, and can reduce the problem of unstable correction quality caused by manual correction.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, in particular to the field of intelligent transportation technology, and specifically to a data processing method, device, electronic device and storage medium. Background Art

[0002] In the roadside perception system, the use effect of the signal light is very important. In actual use, the signal light may be blocked, which will affect the normal use of the signal light.

[0003] In order to implement business processing for traffic lights, for example, online analysis of whether a traffic light is blocked, the captured traffic light image is usually used as a sample image, and the traffic light image is manually annotated to obtain annotated data; wherein the annotated data may include but is not limited to an obstruction attribute for characterizing whether the traffic light is blocked and information for describing the obstruction situation of the traffic light. Summary of the invention

[0004] The present disclosure provides a data processing method, device, electronic device and storage medium.

[0005] According to one aspect of the present disclosure, there is provided a data processing method, comprising:

[0006] Obtaining annotation data for traffic light images;

[0007] If the occlusion attribute in the labeled data indicates that the traffic light is blocked, determining a correction method corresponding to the occlusion attribute according to the target description information in the labeled data; wherein the target description information is used to indicate the occlusion condition of the traffic light;

[0008] The occlusion attribute is corrected according to the determined correction method.

[0009] According to another aspect of the present disclosure, there is provided a data processing device, comprising:

[0010] An acquisition module, used for acquiring annotation data for a traffic light image;

[0011] A determination module, configured to determine, if the occlusion attribute in the labeled data indicates that the traffic light is occluded, a correction method corresponding to the occlusion attribute according to the target description information in the labeled data; wherein the target description information is used to indicate the occlusion condition of the traffic light;

[0012] The correction module is used to correct the occlusion attribute according to the determined correction method.

[0013] The present disclosure also provides an electronic device, comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores instructions that can be executed by the at least one processor. The instructions are executed by the at least one processor so that the at least one processor can perform the above data processing method.

[0017] The present disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the above data processing method.

[0018] The present disclosure also provides a computer program product, including a computer program, which implements the above data processing method when executed by a processor.

[0019] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0021] Figure 1 is a flow chart of a data processing method according to the present disclosure;

[0022] Figure 2 is another flow chart of a data processing method according to the present disclosure;

[0023] Figure 3 is a schematic diagram of a signal light image according to the present disclosure;

[0024] Figure 4 is a schematic diagram of another signal light image according to the present disclosure;

[0025] Figure 5 is a structural schematic diagram of a data processing device according to the present disclosure;

[0026] Figure 6 It is a block diagram of an electronic device used to implement the data processing method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0027] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0028] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0029] In the roadside perception system, the use effect of signal lights is very important, among which signal lights can include but are not limited to traffic lights. In actual use, signal lights may be blocked, which will affect the normal use of signal lights. For example, branches of roadside trees hang down and block signal lights, or passing vehicles or other objects block part of the signal lights because of their height, which will undoubtedly affect the normal use of signal lights.

[0030] In order to realize business processing for traffic lights, for example, online analysis of whether they are blocked, the captured images of traffic lights are usually used as sample images, and the images of traffic lights are manually annotated to obtain annotated data; thus, the sample images and annotated data are used for learning, for example, learning whether the traffic lights are blocked. Through business processing of traffic lights, it is possible to optimize the settings of traffic lights, iterate versions, etc. Among them, the annotated data may include but is not limited to the blockage attributes used to characterize whether the traffic lights are blocked and the information used to describe the blockage situation of the traffic lights. The so-called information used to describe the blockage situation may include the blockage information of the light frame and the blockage information of the light head, etc.

[0031] When manually setting the occlusion attribute, it is usually considered whether the light frame of the traffic light is blocked, so as to set the occlusion attribute of the traffic light to blocked or non-blocked. In this way, the annotation data for the traffic light image may have the following problems: the occlusion attribute indicates that the traffic light is blocked, but it does not actually affect the use of the traffic light, that is, the occlusion attribute cannot completely fit the online judgment of the traffic light occlusion situation, which ultimately leads to the impact on subsequent business processing when the annotation data is used for subsequent business processing. For example: if the edge of the traffic light frame is blocked, or the unlit light of the traffic light is blocked, when manually annotating, although it does not affect the use of the traffic light, the occlusion attribute of the traffic light image containing the traffic light may be set to blocked.

[0032] Based on the above problems, there is a need to correct the occlusion attributes in the labeled data. By correcting the occlusion attributes, the occlusion attributes can be matched with whether the use of the traffic light is affected. That is, when there is an occlusion in the traffic light, if the use of the traffic light is affected, the occlusion attribute is occlusion, and if the use of the traffic light is not affected, the occlusion attribute is non-occlusion. In the related art, since there is no image algorithm that can correct the labeled data, the labeled data is usually corrected manually. However, the manual correction of the occlusion attributes in the labeled data undoubtedly has the problem of low correction efficiency. Based on the above content, it can be seen that how to improve the efficiency of correcting the occlusion attributes in the labeled data is a problem that needs to be solved urgently.

[0033] In order to solve the above problems, the present disclosure provides a data processing method, device, electronic device and storage medium.

[0034] The following first introduces a data processing method provided by an embodiment of the present disclosure.

[0035] A data processing method provided by an embodiment of the present disclosure can be applied to an electronic device. In a specific application, the electronic device can be a server or a terminal device, which is reasonable.

[0036] Specifically, the execution subject of the data processing method may be a data processing device. Exemplarily, when the data processing method is applied to a terminal device, the data processing device may be a functional software running in the terminal device. Exemplarily, when the data processing method is applied to a server, the data processing device may be a computer program running in the server, and the computer program may be used for data processing.

[0037] A data processing method provided in an embodiment of the present disclosure may include the following steps:

[0038] Obtaining annotation data for traffic light images;

[0039] If the occlusion attribute in the labeled data indicates that the traffic light is blocked, determining a correction method corresponding to the occlusion attribute according to the target description information in the labeled data; wherein the target description information is used to indicate the occlusion condition of the traffic light;

[0040] The occlusion attribute is corrected according to the determined correction method.

[0041] In this solution, whether the actual use of the traffic light is affected is considered to be correlated with the traffic light occlusion situation. Therefore, the target description information can be relied upon in the correction process. Moreover, the credibility of the information included in the labeled data under different traffic light occlusion situations is different. Therefore, in the correction process, the correction method corresponding to the occlusion attribute can be determined based on the target description information, so as to correct the occlusion attribute. It can be seen that this solution uses the correction method determined by the target description information to correct the occlusion attribute, which can improve the efficiency of correcting the occlusion attribute in the labeled data compared with the manual correction method.

[0042] In addition, this solution can reduce the problem of unstable correction quality caused by manual correction.

[0043] A data processing method provided by the present disclosure is introduced below in conjunction with the accompanying drawings.

[0044] like Figure 1 As shown, a data processing method provided by an embodiment of the present disclosure may include the following steps:

[0045] S101, obtaining annotation data for a traffic light image;

[0046] Among them, the annotation data for the traffic light image can be manually annotated on the traffic light image, and the annotation data can include an occlusion result used to characterize whether the traffic light is blocked, as well as target description information of the traffic light. The target description information may include light frame occlusion information, lamp head lighting status, occlusion information of the lit lamp head, shape information, position information, etc.

[0047] In addition, it can be understood that the signal light image is any image obtained by photographing the signal light and having annotation data. In a specific application, the signal light image can be bound and saved with the annotation data it has, so that the electronic device can first determine the signal light image, and then obtain the annotation data for the signal light image; wherein, there can be multiple ways to determine the signal light image, exemplarily, it can be specified from a signal light image list, or an image can be input through an interactive interface to obtain the signal light image, and so on. Of course, the electronic device can also obtain the annotation data from a predetermined database for storing the annotation data of each signal light image, so as to obtain the annotation data for the signal light image. It should be emphasized that the embodiments of the present disclosure do not limit the specific implementation method of obtaining the annotation data for the signal light image.

[0048] Furthermore, the signal lights in the signal light image in the present disclosure may include traffic lights on the road for guiding traffic, and of course may also include other signal lights for warning or guiding. In the case where the signal light is a traffic light, the signal light may be a signal light at any intersection, and the present disclosure does not limit its position, shape, or size.

[0049] In addition, the traffic light images mentioned in this embodiment are related to actual usage scenarios, and the collection, storage, use, processing, transmission, provision and disclosure of traffic light images comply with relevant laws and regulations and do not violate public order and good morals.

[0050] S102, if the occlusion attribute in the annotated data indicates that the traffic light is occluded, determining a correction method corresponding to the occlusion attribute according to target description information in the annotated data; wherein the target description information is used to indicate the occlusion condition of the traffic light;

[0051] When the occlusion attribute in the annotated data indicates that the traffic light is obscured, it may not actually affect the use of the traffic light, that is, the occlusion attribute cannot completely fit the online judgment of the traffic light occlusion situation. Therefore, in this step, when the occlusion attribute in the acquired annotated data indicates that the traffic light is obscured, the process of correcting the occlusion attribute continues, that is, the correction method corresponding to the occlusion attribute is determined according to the target description information in the annotated data. It can be understood that if the occlusion attribute in the acquired annotated data indicates that the traffic light is not obscured, that is, the traffic light is not obscured, then the correction process for the acquired annotated data can be terminated.

[0052] Furthermore, the inventors found in their research that whether the actual use of traffic lights is affected is correlated with the traffic light occlusion situation, and the credibility of the information included in the labeled data under different traffic light occlusion situations is different. Therefore, in the correction process, the target description information can be used as a reference, and the correction method corresponding to the occlusion attribute can be determined based on the target description information, so as to subsequently correct the occlusion attribute. It can be understood that different target description information can correspond to different correction methods. The so-called correction method corresponding to the occlusion attribute is to select the correction method to be used from multiple correction methods, so as to ensure that the correction result is more accurate.

[0053] Exemplarily, the occlusion result may be characterized as occlusion or non-occlusion.

[0054] Exemplarily, the target description information may include: the size of the light frame occlusion area, the occlusion information of the lit lamp head, the shape of the lit lamp head and / or its positional relationship with the light frame, etc., which will be replaced by the first description information, the second description information and the third description information below. The present disclosure does not impose too many restrictions on the specific content and specific content name of the target description information.

[0055] In addition, multiple correction methods can be set based on other information in the annotation data. Other information may include but is not limited to: one or more of the lighting status of the lamp head, the occlusion information of the lit lamp head, shape information, and position information. The specific information to be used can be selected according to the credibility of other descriptive information. Exemplary: For the case of high credibility, only one kind of information can be used for correction processing, such as: the occlusion attribute or shape information of the lit lamp head; and the corresponding correction method may include: if the one kind of information represents that the use of the signal light is not affected, for example: the lit lamp heads are not blocked or the shapes are all set to be known, then the occlusion attribute is set to non-occlusion, otherwise, the correction operation is rejected. Exemplary, for the case of credibility, multiple kinds of information can be used for correction processing, such as: the occlusion information and shape information of the lit lamp head; and the corresponding correction method may include: if the multiple kinds of information simultaneously represent that the use of the signal light is not affected, for example: the lit lamp heads are not blocked and the shapes are all set to be known, then the occlusion attribute is set to non-occlusion, otherwise, the correction operation is rejected. Other setting methods of the various correction methods are described in detail in another embodiment of the present disclosure and will not be elaborated on here.

[0056] S103: Correct the occlusion attribute according to the determined correction method.

[0057] After the correction method is determined, the correction process can be performed according to the step content indicated by the determined correction method. It should be noted that since the occlusion attribute may not match the actual use of the traffic light, or may match the actual use of the traffic light, there may be two correction results obtained by correcting the occlusion attribute: the occlusion attribute is changed to non-occlusion, or the occlusion attribute remains as occlusion.

[0058] Regarding the specific implementation method of correcting the occlusion attribute according to the determined correction method, another embodiment will be introduced later in combination with the specific correction method.

[0059] In this solution, whether the actual use of the traffic light is affected is considered to be correlated with the traffic light occlusion situation. Therefore, the target description information can be relied upon in the correction process. Moreover, the credibility of the information included in the labeled data under different traffic light occlusion situations is different. Therefore, in the correction process, the correction method corresponding to the occlusion attribute can be determined based on the target description information, so as to correct the occlusion attribute. It can be seen that this solution uses the correction method determined by the target description information to correct the occlusion attribute, which can improve the efficiency of correcting the occlusion attribute in the labeled data compared with the manual correction method.

[0060] In addition, this solution can reduce the problem of unstable correction quality caused by manual correction.

[0061] Optionally, the target description information may include first description information for characterizing the size of the shielding area of ​​the light frame; under the premise that the target description information includes the first description information, Figure 2 As shown, in another embodiment of the present disclosure, a data processing method is provided, which may include steps S201-S203:

[0062] S201, obtaining annotation data for a traffic light image;

[0063] In this embodiment, step S201 is the same as step S101 in the above embodiment, and will not be described in detail here.

[0064] S202, if the occlusion attribute in the annotated data indicates that the traffic light is occluded, determine a correction method corresponding to the occlusion attribute according to a magnitude relationship between the first description information in the annotated data and a predetermined threshold value; wherein the target description information is used to indicate the occlusion condition of the traffic light;

[0065] Since the size of the blocked area is an intuitive reflection of the traffic light blockage situation, and the credibility of the information included in the annotation data is different under different blockage area sizes, therefore, when the blockage attribute characterizes that the traffic light is blocked, the correction method corresponding to the blockage attribute can be determined based on the size relationship between the first descriptive information in the annotation data and the predetermined threshold, that is, a correction method is selected from multiple correction methods.

[0066] There may be multiple representation forms of the first description information. Exemplarily, the representation form of the first description information may be a level, for example: the value of the first description information may include multiple levels: level 1, level 2, level 3, wherein from level 1 to level 3, the size of the light frame obstruction area represented gradually increases. Exemplarily, the representation form of the first description information may be a percentage, that is, the percentage of the light frame obstruction area to the entire light frame area.

[0067] It can be understood that the representation form of the predetermined threshold is the same as the representation form of the first description information, wherein the predetermined threshold can be used to distinguish whether the current light frame shading area is high or low, and the present disclosure does not make specific limitations on the predetermined threshold.

[0068] S203: Correct the occlusion attribute according to the determined correction method.

[0069] In this embodiment, step S201 is the same as step S101 in the above embodiment, and will not be described in detail here.

[0070] It can be seen that in this embodiment, compared with the manual correction method, the efficiency of correcting the occlusion attribute in the labeled data can be improved. In addition, according to the size relationship between the first description information in the labeled data and the predetermined threshold, the correction method corresponding to the occlusion attribute is determined, so that the correction method determination process can be more convenient.

[0071] Optionally, in another embodiment of the present disclosure, the target description information may further include: second description information for characterizing whether the lighted lamp head of the signal lamp is blocked, and third description information for characterizing the shape of the lighted lamp head and / or the positional relationship with the lamp frame;

[0072] Determining the correction method corresponding to the occlusion attribute according to the magnitude relationship between the first description information in the annotated data and the predetermined threshold may include the following steps A1-A2:

[0073] Step A1, if the area size represented by the first description information in the annotation data is lower than the area size represented by the predetermined threshold, determining the first correction method as the correction method corresponding to the occlusion attribute;

[0074] Step A2: if the area size represented by the first description information in the annotation data is not less than the area size represented by the predetermined threshold, determining the second correction method as the correction method corresponding to the occlusion attribute;

[0075] The first correction method is a correction method based on the second description information, and the second correction method is a correction method based on the second description information and the third description information.

[0076] The inventor has found through research that whether the lit lamp head represented by the second description information is blocked is the key to determining whether the signal light can be used.

[0077] Furthermore, if the area size represented by the first descriptive information in the annotation data is lower than the area size represented by the predetermined threshold, that is, the area blocked by the lamp frame is small, the annotation personnel are not easily disturbed by the blockage and can usually mark the blockage situation more accurately, so that the credibility of the manually marked second descriptive information is higher. At this time, corrections can be made based solely on the second descriptive information, that is, corrections can be made based on whether the light head of the traffic light is blocked.

[0078] If the area represented by the first description information in the annotation data is not less than the area represented by the predetermined threshold, that is, the light frame obstructs a large area, the annotation personnel are usually easily disturbed by the obstruction and are prone to errors when annotating the obstruction, making the credibility of the manually annotated second description information relatively low. Then, the third description information can be added on the basis of the second description information, and corrections can be made based on the second and third description information. In this way, the accuracy of the correction can be guaranteed by considering a richer basis for judgment.

[0079] For example, if the representation form of the first description information is a level, and specifically includes three levels: level 1, level 2 and level 3, and the occlusion area gradually increases from level 1 to level 3; assuming that the predetermined threshold is set at level 3; if the value of the first description information is lower than level 3, it indicates that the area represented by the first description information is lower than the area represented by the predetermined threshold. At this time, the first correction method can be used for correction analysis, that is, correction is performed only based on the second description information;

[0080] If the value of the second description information is not lower than level 3, it indicates that the area represented by the first description information is not lower than the area represented by the predetermined threshold. At this time, the second correction method can be used for correction analysis, that is, correction is performed based on the second description information and the third description information.

[0081] For example, if the representation form of the first description information is a percentage; assuming that the predetermined threshold is set to 70%; if the value of the first description information is lower than 70%, it indicates that the area represented by the first description information is lower than the area represented by the predetermined threshold. At this time, the first correction method can be used for correction analysis, that is, correction is performed only based on the second description information;

[0082] If the value of the second description information is not less than 70%, it indicates that the area represented by the first description information is not less than the area represented by the predetermined threshold. At this time, the second correction method can be used for correction analysis, that is, correction is performed based on the second description information and the third description information.

[0083] It can be seen that in this embodiment, the first correction method or the second correction method is adopted by utilizing the relationship between the area size represented by the first description information in the annotation data and the area size represented by the predetermined threshold, so that the correction processing of the occlusion attribute is more accurate.

[0084] Optionally, if the first correction method is a correction method based on the second description information, in another embodiment of the present disclosure, the correction processing of the occlusion attribute according to the determined correction method includes B1-B2:

[0085] Step B1, if the determined correction method is the first correction method, determining whether the second description information indicates that the lit lamp head is not blocked;

[0086] Step B2: if the judgment result shows that the lighted lamp heads of the signal lights are not blocked, the blocking attribute is corrected to non-blocked; otherwise, the correction operation is rejected.

[0087] If the determined correction method is the first correction method, the occlusion attribute can be corrected according to the correction steps given in the first correction method. The correction steps here are: determine whether the second description information indicates that the lit lamp head is non-occluded, and if the judgment result shows that the lit lamp heads of the signal light are all non-occluded, then correct the occlusion attribute to non-occluded; otherwise, reject the correction operation.

[0088] It can be understood that if the lit lamp heads of the signal lights are all unobstructed, it can be shown that the lit lamp heads of the signal lights are visible, and the signal lights are not affected during actual use. Therefore, the obstruction attribute can be corrected to non-obstruction; if the lit lamp heads of the signal lights are obstructed, it can be shown that the signal lights are affected during actual use. Therefore, the obstruction attribute is obstruction, which is reasonable, and the correction operation can be rejected, that is, the obstruction attribute can be kept as obstruction.

[0089] For example, if the first descriptive information level in a traffic light annotation data is 1, the correction method is determined to be the first correction method. At this time, based on the first correction method, it is necessary to perform an occlusion judgment on the lit lamp head of the traffic light. If all the lit lamp heads are not obstructed, the correction processing is performed to correct the occlusion attribute of the traffic light to non-obstructed; if a lit lamp head of the traffic light is obstructed, the correction processing of the occlusion attribute of the traffic light is refused.

[0090] It can be seen that in this solution, it is determined whether the lighted lamp head of the signal light is not blocked, and different operations are performed according to different judgment results, so that the correction process is simple and effective.

[0091] Optionally, if the second correction method is a correction method based on the second description information and the third description information, the third description information specifically includes: first information for characterizing the shape of the lit lamp head, second information for characterizing the position information of the lit lamp head, and third information for characterizing the position information of the lamp frame;

[0092] Correspondingly, in another embodiment of the present disclosure, the correction processing of the occlusion attribute according to the determined correction method includes steps C1-C2:

[0093] Step C1, if the determined correction method is the second correction method, judging whether the lit lamp head is not blocked based on the second description information, judging whether the shape of the lit lamp head is not an unknown shape based on the first information, and judging whether the position of the lit lamp head is located within the lamp frame based on the second information and the third information;

[0094] Step C2: if the judgment results are all yes, the occlusion attribute is corrected to non-occlusion; otherwise, the correction operation is rejected.

[0095] If the determined correction method is the second correction method, the occlusion attribute can be corrected according to the correction steps given by the second correction method. The correction steps here are: based on the second description information, determine whether the lit lamp head is non-occluded, and based on the first information, determine whether the shape of the lit lamp head is not an unknown shape, based on the second information and the third information, determine whether the position of the lit lamp head is located within the lamp frame, and if the judgment results are all yes, correct the occlusion attribute to non-occlusion, otherwise, reject the correction operation.

[0096] It can be understood that although the credibility of various types of information in the annotation information is low when the obstruction area is large, if three types of information all indicate that the lit lamp head of the traffic light is visible, that is, the lit lamp head is not obstructed, and the shape of the lit lamp head is not an unknown shape, and the position of the lit lamp head is located in the lamp frame, it means that the traffic light is not affected during actual use. Therefore, the obstruction attribute can be changed to non-obstruction; otherwise, the correction operation can be rejected, that is, the obstruction attribute can be kept as obstruction.

[0097] For example, if the first descriptive information level in the annotation data of a traffic light is 3, the correction method is determined to be the second correction method. At this time, based on the second correction method, it is necessary to perform an occlusion judgment on the lit lamp head of the traffic light, and judge whether the shape of the lit lamp head does not belong to an unknown shape based on the first information. Based on the second information and the third information, judge whether the position of the lit lamp head is located in the lamp frame. If all the lit lamp heads are not obstructed, the shape of the lit lamp heads does not belong to an unknown shape, and the position of the lit lamp heads is located in the lamp frame, then the correction processing is performed to correct the occlusion attribute of the traffic light to non-obstruction; if the lit lamp head of the traffic light is obstructed, the shape of the lit lamp head does not belong to an unknown shape, and the position of the lit lamp head is located in the lamp frame, then the correction processing of the occlusion attribute of the traffic light is refused.

[0098] In this solution, the three types of information are combined to determine whether the occlusion attribute needs to be corrected, and different operations are performed according to different judgment results, which can ensure the effectiveness and accuracy of the correction process.

[0099] In order to better explain the content of the embodiments of the present disclosure, Figure 3 and Figure 4 A specific example of the present disclosure is described.

[0100] Taking a traffic light image taken by a roadside camera as an example, a specific example of the present disclosure may include the following content:

[0101] Step D1, obtaining the annotation data of the traffic light image.

[0102] Among them, traffic_light_occluded_attribute is used as the field identifier of the occlusion attribute, where non-occlusion is 0 and occlusion is 1;

[0103] Traffic_light_occluded_degree is used as the field identifier of the first description information, which specifically includes level 1 to level 3. From level 1 to level 3, the occlusion area represented gradually increases; for example: Figure 3 The occlusion in can be marked with level 1, and Figure 4 For the occlusion in the image, you can use level 3 identification;

[0104] light_item_occluded_attribute is used as the field identifier of the second description information, where unoccluded is 0 and occluded is 1;

[0105] light_item_shape is used as the field identifier of the first information, where 0 indicates unknown and is greater than 1 in other cases;

[0106] light_item_x, light_item_y, light_item_w, and light_item_h are used as field identifiers of the second information;

[0107] Traffic_light_x, traffic_light_y, traffic_light_w, and traffic_light_h are used as field identifiers of the third information; and based on the second information and the third information, it can be calculated whether the lit lamp head is within the light frame, which is expressed as light_item_out_traffic_light, where 0 indicates within the light frame and 1 indicates outside the light frame.

[0108] Furthermore, the second descriptive information, the first information, and whether the calculated lamp head position is within the lamp frame can be stored in three lists respectively: a list light_occ_list for storing the second descriptive information of the lit lamp head, a list light_shape_list for storing the first information of the lit lamp head, and a list light_out_tl_list for storing whether the lit lamp head is within the lamp frame.

[0109] Step D2, if traffic_light_occluded_attribute is 0, skip it, and execute step D3 only if it is greater than 0.

[0110] Step D3, determining whether the value of the first description information traffic_light_occluded_degree is less than level 3, if it is less than level 3, executing step D4; if it is equal to level 3, executing step D5.

[0111] Step D4, when it is determined that the second description information light_occ_list of all lit lamp heads exists and the maximum value of the second description information light_occ_list of all lit lamp heads is equal to 0, it means that all lit lamp heads are not blocked, then the blocking attribute traffic_light_occluded_attribute is modified to 0, otherwise, the correction operation is rejected.

[0112] Step D5, if the following specified conditions are met, the occlusion attribute traffic_light_occluded_attribute is modified to 0, otherwise, the correction operation is rejected:

[0113] The specified conditions include:

[0114] The second description information light_occ_list of all the lit lamp heads exists, and the maximum value of the second description information light_occ_list of all the lit lamp heads is equal to 0;

[0115] And the first information light_shape_list of all the lit lamp heads exists, and the minimum value of the first information light_shape_list of all the lit lamp heads is greater than 0;

[0116] In addition, the light_out_tl_list, which indicates whether all the positions of the lit lamp heads are within the lamp frame, exists, and the maximum value of the light_out_tl_list, which indicates whether all the positions of the lit lamp heads are within the lamp frame, is equal to 0.

[0117] If the specified conditions are met, it means that all the lit lamp heads are not blocked, all the lit lamp heads have shapes, and all the lamp heads are within the lamp frame.

[0118] In this solution, whether the actual use of the traffic light is affected is considered to be correlated with the traffic light occlusion situation. Therefore, the target description information can be relied upon in the correction process. Moreover, the credibility of the information included in the labeled data under different traffic light occlusion situations is different. Therefore, in the correction process, the correction method corresponding to the occlusion attribute can be determined based on the target description information, so as to correct the occlusion attribute. It can be seen that this solution uses the correction method determined by the target description information to correct the occlusion attribute, which can improve the efficiency of correcting the occlusion attribute in the labeled data compared with the manual correction method.

[0119] In addition, this solution can reduce the problem of unstable correction quality caused by manual correction.

[0120] Based on the above method embodiment, the present disclosure also provides a data processing device, such as Figure 5 As shown, the device comprises:

[0121] An acquisition module 510, used to acquire annotation data for a traffic light image;

[0122] Determining module 520, if the occlusion attribute in the labeled data indicates that the traffic light is blocked, determining a correction method corresponding to the occlusion attribute according to the target description information in the labeled data; wherein the target description information is used to indicate the occlusion condition of the traffic light;

[0123] The correction module 530 is used to correct the occlusion attribute according to the determined correction method.

[0124] Optionally, the target description information includes: first description information for characterizing the size of the light frame shielding area;

[0125] The determining module comprises:

[0126] The determination submodule is used to determine the correction method corresponding to the occlusion attribute according to the size relationship between the first description information in the annotation data and a predetermined threshold.

[0127] Optionally, the target description information further includes: second description information for characterizing whether the lighted lamp head of the signal lamp is blocked, and third description information for characterizing the shape of the lighted lamp head and / or the positional relationship with the lamp frame;

[0128] The determination submodule comprises:

[0129] A determination unit, configured to determine the first correction method as the correction method corresponding to the occlusion attribute if the area size represented by the first description information in the annotation data is lower than the area size represented by the predetermined threshold; and to determine the second correction method as the correction method corresponding to the occlusion attribute if the area size represented by the first description information in the annotation data is not lower than the area size represented by the predetermined threshold;

[0130] The first correction method is a correction method based on the second description information, and the second correction method is a correction method based on the second description information and the third description information.

[0131] Optionally, the correction module includes:

[0132] A first judgment submodule, configured to judge whether the second description information indicates that the lit lamp head is not blocked if the determined correction method is the first correction method;

[0133] If the judgment result shows that the lighted lamp heads of the signal lights are not blocked, the blocking attribute is corrected to non-blocked, otherwise, the correction operation is rejected.

[0134] Optionally, the third description information specifically includes: first information for characterizing the shape of the lit lamp head, second information for characterizing the position information of the lit lamp head, and third information for characterizing the position information of the lamp frame;

[0135] Correction modules, including:

[0136] The second judgment submodule is used for, if the determined correction method is the second correction method,

[0137] Determine whether the lit lamp head is not blocked based on the second description information, determine whether the shape of the lit lamp head is not an unknown shape based on the first information, and determine whether the position of the lit lamp head is within the lamp frame based on the second information and the third information;

[0138] If the judgment results are all yes, the occlusion attribute is corrected to non-occlusion, otherwise, the correction operation is rejected.

[0139] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0140] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0141] The present disclosure provides an electronic device, including:

[0142] at least one processor; and

[0143] a memory communicatively connected to at least one processor; wherein,

[0144] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the data processing method.

[0145] Among them, the present disclosure provides a readable storage medium, in which computer instructions are stored, and the computer instructions are used to enable a computer to execute a data processing method.

[0146] Among them, a computer program product provided by the present disclosure includes a computer program, and the computer program implements the data processing method when executed by a processor.

[0147] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0148] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0149] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0150] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as data processing methods. For example, in some embodiments, the data processing method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the data processing method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the data processing method in any other appropriate manner (e.g., by means of firmware).

[0151] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0152] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0153] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0154] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0155] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0156] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0157] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0158] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A data processing method, comprising: Obtaining annotation data for traffic light images; If the occlusion attribute in the labeled data indicates that the traffic light is blocked, determining a correction method corresponding to the occlusion attribute according to the target description information in the labeled data; wherein the target description information is used to indicate the occlusion condition of the traffic light; According to the determined correction method, the occlusion attribute is corrected; wherein the correction result obtained by the correction process includes that the occlusion attribute is changed to non-occlusion, or the occlusion attribute remains as occlusion.

2. The method according to claim 1, wherein: The target description information includes: first description information for characterizing the size of the light frame shielding area; The determining, according to the target description information in the labeled data, a correction method corresponding to the occlusion attribute includes: A correction method corresponding to the occlusion attribute is determined according to a magnitude relationship between the first description information in the annotation data and a predetermined threshold.

3. The method according to claim 2, wherein: The target description information further includes: second description information for characterizing whether the lighted lamp cap of the signal lamp is blocked, and third description information for characterizing the shape of the lighted lamp cap and / or the positional relationship with the lamp frame; The determining, according to the magnitude relationship between the first description information in the annotated data and a predetermined threshold, a correction method corresponding to the occlusion attribute includes: If the area size represented by the first description information in the annotation data is lower than the area size represented by the predetermined threshold, the first correction method is determined as the correction method corresponding to the occlusion attribute; If the area size represented by the first description information in the annotation data is not less than the area size represented by the predetermined threshold, the second correction method is determined as the correction method corresponding to the occlusion attribute; The first correction method is a correction method based on the second description information, and the second correction method is a correction method based on the second description information and the third description information.

4. The method according to claim 3, wherein: The correcting process of the occlusion attribute according to the determined correction method includes: If the determined correction method is the first correction method, determining whether the second description information indicates that the lit lamp head is not blocked; If the judgment result shows that the lighted lamp heads of the signal lights are not blocked, the blocking attribute is corrected to non-blocked, otherwise, the correction operation is rejected.

5. The method according to claim 3, wherein: The third description information specifically includes: first information for characterizing the shape of the lit lamp head, second information for characterizing the position information of the lit lamp head, and third information for characterizing the position information of the lamp frame; According to the determined correction method, the occlusion attribute is corrected, including: If the determined correction method is the second correction method, determining whether the lit lamp head is not blocked based on the second description information, determining whether the shape of the lit lamp head is not an unknown shape based on the first information, and determining whether the position of the lit lamp head is within the lamp frame based on the second information and the third information; If the judgment results are all yes, the occlusion attribute is corrected to non-occlusion, otherwise, the correction operation is rejected.

6. A data processing device, comprising: An acquisition module, used for acquiring annotation data for a traffic light image; A determination module, configured to determine, if the occlusion attribute in the labeled data indicates that the traffic light is occluded, a correction method corresponding to the occlusion attribute according to the target description information in the labeled data; wherein the target description information is used to indicate the occlusion condition of the traffic light; The correction module is used to perform correction processing on the occlusion attribute according to the determined correction method; wherein the correction result obtained by the correction processing includes that the occlusion attribute is changed to non-occlusion, or the occlusion attribute remains as occlusion.

7. The device according to claim 6, wherein: The target description information includes: first description information for characterizing the size of the light frame shielding area; The determining module comprises: The determination submodule is used to determine the correction method corresponding to the occlusion attribute according to the size relationship between the first description information in the annotation data and a predetermined threshold.

8. The device according to claim 7, wherein: The target description information further includes: second description information for characterizing whether the lighted lamp cap of the signal lamp is blocked, and third description information for characterizing the shape of the lighted lamp cap and / or the positional relationship with the lamp frame; The determination submodule comprises: A determination unit, configured to determine the first correction method as the correction method corresponding to the occlusion attribute if the area size represented by the first description information in the annotation data is lower than the area size represented by the predetermined threshold; and to determine the second correction method as the correction method corresponding to the occlusion attribute if the area size represented by the first description information in the annotation data is not lower than the area size represented by the predetermined threshold; The first correction method is a correction method based on the second description information, and the second correction method is a correction method based on the second description information and the third description information.

9. The device according to claim 8, wherein: The correction module comprises: The first judgment submodule is used to judge whether the second description information represents that the lit lamp head is not obstructed if the determined correction method is the first correction method; if the judgment result shows that the lit lamp heads of the signal lights are not obstructed, the obstruction attribute is corrected to non-obstructed, otherwise, the correction operation is rejected.

10. The device according to claim 8, wherein: The third description information specifically includes: first information for characterizing the shape of the lit lamp head, second information for characterizing the position information of the lit lamp head, and third information for characterizing the position information of the lamp frame; The correction module comprises: A second judgment submodule is used for judging whether the lit lamp head is not blocked based on the second description information if the determined correction method is the second correction method, judging whether the shape of the lit lamp head is not an unknown shape based on the first information, and judging whether the position of the lit lamp head is located within the lamp frame based on the second information and the third information; If the judgment results are all yes, the occlusion attribute is corrected to non-occlusion, otherwise, the correction operation is rejected.

11. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

12. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-5.

13. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 5.

Images