Picture audit method and apparatus, electronic device, and storage medium

Abstract

The present disclosure relates to a picture audit method and apparatus, an electronic device, and a storage medium. The picture audit method comprises: acquiring pictures to be audited, wherein the number of said pictures is at least two; using preset algorithms to calculate the similarities between said pictures, wherein the preset algorithms include at least two similarity calculation algorithms; and on the basis of the plurality of similarities between said pictures, determining whether said pictures are similar pictures.

Description

Picture auditing method and device, electronic equipment and storage medium

[0001] Cross-reference to Related Applications

[0002] The present application claims priority to Chinese Patent Application No. 202411100007.0, filed on August 12, 2024, the contents of which are incorporated herein by reference in its entirety. TECHNICAL FIELD

[0003] The present disclosure relates to the technical field of image processing, and in particular to a picture auditing method and device, electronic equipment and storage medium. BACKGROUND

[0004] In related technologies, maintenance work photo auditing is an important basis for evaluating the quality of maintenance personnel. Maintenance work photos refer to check-in photos, equipment photos and other proof materials uploaded by maintenance personnel during work execution. Since maintenance work photos may be adjusted by filters, cropping, picture rotation and other methods for the same group of photos as proof materials for other work orders, and in fact no work has been performed. Therefore, it is necessary to audit maintenance work photos, and currently, manual auditing is usually relied on. However, due to the large number of maintenance work photos, manual auditing is prone to judgment errors due to fatigue and other reasons, resulting in low auditing efficiency, low accuracy and completeness of the auditing results. SUMMARY

[0005] The present disclosure provides a picture auditing method, device, electronic equipment and storage medium to at least solve the technical problems of low auditing efficiency, low accuracy and completeness of the auditing results in related technologies. The technical solutions of the present disclosure are as follows:

[0006] According to a first aspect of an embodiment of the present disclosure, a picture auditing method is provided, comprising:

[0007] Obtaining a picture to be audited; the picture to be audited is at least two pictures;

[0008] Calculating the similarity between the pictures to be audited using a preset algorithm; wherein the preset algorithm includes at least two similarity calculation algorithms;

[0009] Determining whether the pictures to be audited are similar pictures according to a plurality of similarities between the pictures to be audited.

[0010] In a possible implementation, before calculating the similarity between the pictures to be audited using the preset algorithm, the method further comprises:

[0011] determine preset target information in each of the to-be-audited pictures through a preset AI recognition model, wherein the preset target information comprises preset targets and the number of each of the preset targets;

[0012] In a case where the preset target information in the to-be-audited pictures is different, determine that the to-be-audited pictures are dissimilar pictures.

[0013] The calculating the similarity between the to-be-audited pictures through a preset algorithm comprises:

[0014] In a case where the preset target information in the to-be-audited pictures is different, determine that the to-be-audited pictures are dissimilar pictures.

[0015] In a possible implementation, the similarity calculation algorithm comprises a mean hash algorithm, a difference hash algorithm, and a perceptual hash algorithm.

[0016] The calculating the similarity between the to-be-audited pictures through a preset algorithm comprises:

[0017] The calculating the similarity between the to-be-audited pictures through a preset algorithm comprises:

[0018] The determining whether the to-be-audited pictures are similar pictures according to the plurality of similarities between the to-be-audited pictures comprises:

[0019] The determining whether the to-be-audited pictures are similar pictures according to the plurality of similarities between the to-be-audited pictures comprises:

[0020] In a possible implementation, the determining whether the to-be-audited pictures are similar pictures according to the plurality of similarities between the to-be-audited pictures comprises:

[0021] Selecting the minimum similarity among the first similarity, the second similarity, and the third similarity.

[0022] In a case where the minimum similarity is less than a first preset similarity threshold, determine that the to-be-audited pictures are dissimilar pictures.

[0023] In a possible implementation, the picture auditing method further comprises:

[0024] In a case where the minimum similarity is greater than or equal to the first preset similarity threshold, determine whether the minimum similarity is greater than a second preset similarity threshold.

[0025] In a case where the minimum similarity is greater than the second preset similarity threshold, determine that the to-be-audited pictures are similar pictures.

[0026] In a possible implementation, the similarity calculation algorithm further includes a scale-invariant feature transform algorithm.

[0027] The picture auditing method further includes:

[0028] In a case where the minimum similarity is less than the second preset similarity threshold, a fourth similarity between the pictures to be audited is calculated by using the scale-invariant feature transform algorithm.

[0029] In a case where the fourth similarity is greater than the second preset similarity threshold, the pictures to be audited are determined as similar pictures.

[0030] In a case where the fourth similarity is less than or equal to the second preset similarity threshold, the pictures to be audited are determined as dissimilar pictures.

[0031] According to a second aspect of an embodiment of the present disclosure, a picture auditing apparatus is provided, including:

[0032] a picture receiving module configured to receive pictures to be audited; the pictures to be audited are at least two;

[0033] a calculation module configured to calculate similarities between the pictures to be audited by using a preset algorithm; the preset algorithm includes at least two similarity calculation algorithms;

[0034] an auditing module configured to determine whether the pictures to be audited are similar pictures according to the similarities between the pictures to be audited.

[0035] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, including:

[0036] a processor;

[0037] a memory configured to store instructions executable by the processor;

[0038] The processor is configured to execute the instructions to implement the picture auditing method according to any one of the first aspect.

[0039] According to a fourth aspect of an embodiment of the present disclosure, a storage medium is provided, when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the picture auditing method according to any one of the first aspect.

[0040] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, including a computer program, when the computer program is executed by a processor, the picture auditing method according to any one of the first aspect is implemented.

[0041] The technical solutions provided by the embodiments of the present disclosure at least have the following beneficial effects:

[0042] In the embodiments of the present disclosure, a picture to be audited is obtained; the picture to be audited is at least two; a preset algorithm is used to calculate the similarity between the pictures to be audited; wherein the preset algorithm includes at least two similarity calculation algorithms; and whether the pictures to be audited are similar pictures is determined according to the plurality of similarities between the pictures to be audited. In this way, the similarity between the pictures to be audited can be calculated by multiple algorithms, and whether the pictures to be audited are similar pictures can be determined by the similarity. In this way, the automatic auditing processing of the maintenance work pictures can be realized by multiple similarity algorithms. Compared with manual auditing processing, not only can the judgment errors easily occurred in manual auditing be effectively avoided, the picture auditing efficiency can be effectively improved, and the accuracy and integrity of the picture auditing result can be improved; but also the picture auditing method can be applied to the case where the number of maintenance work pictures is large, and the applicability of the picture auditing method can be improved.

[0043] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0044] The accompanying drawings incorporated in the specification hereof and forming a part thereof illustrate embodiments consistent with the present disclosure and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0045] FIG. 1 is a flowchart of a picture auditing method according to an embodiment of the present disclosure.

[0046] FIG. 2 is a flowchart of a picture auditing method according to an embodiment of the present disclosure.

[0047] FIG. 3 is a block diagram of an execution system of a picture auditing method according to an embodiment of the present disclosure.

[0048] FIG. 4 is a block diagram of a picture auditing device according to an embodiment of the present disclosure.

[0049] FIG. 5 is a block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0050] In order for those skilled in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the accompanying drawings.

[0051] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Rather, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0052] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data for analysis, stored data, displayed data, etc.) and signals involved in the present disclosure are authorized by the user or fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0053] As can be known from the background, in the related art, the maintenance work photo audit is an important basis for judging the quality of the maintenance personnel. Through the audit result, the city company can determine whether the maintenance personnel completes the related maintenance work on time and with quality, and whether the maintenance personnel timely and accurately solves various problems fed back by the customer. The maintenance work photo refers to the check-in photo, equipment photo and the like uploaded by the maintenance personnel through a mobile phone application program (APP) as proof materials of work execution in the process of executing the work. However, there is a fraud risk in the uploading of the work photo: the maintenance personnel adjusts the same group of photos through filters, cropping, picture rotation and the like as proof materials of other work orders, but actually does not execute the work. The maintenance administrator needs to spend a lot of effort to audit the work photo. At present, the artificial audit processing is mainly relied on, and the check-in photo of the work order is checked every month in the maintenance management system to determine whether there is a similar or same photo. At present, the work order data of the maintenance system reaches more than 300,000 every month, and there are 10 photos of the work order on average, so the total amount of the work order photos every month exceeds 3,000,000, and the artificial auditing is time-consuming and costly. At the same time, the false photo can appear in the work order of different regions and different times, or even the same false photo appears in the cross-month work order, and the above situations bring great difficulty to the artificial audit: long-time uninterrupted auditing is easy to cause judgment errors due to fatigue; the division of different people cannot identify all the photos that have appeared, and there is a problem of low accuracy and completeness of the audit. That is, the current scheme has at least the following problems:

[0054] 1. The cost of human resources is high, and manual auditing is time-consuming and inefficient. 2. The accuracy of auditing is low. There are mainly three types of fake photos in existing work: adjusting the same photo using filters, cropping, rotating, etc.; taking photos of the same scene from different angles (the same scene refers to the same maintenance personnel, the same clothing, and the same location); fake photos appearing on different types of work orders at different times. Manual auditing is prone to errors due to fatigue, and the number of fake photos is large and the time span may be large, making it difficult for auditors to identify. 3. The recognition integrity is low. Due to the large number of existing work orders, manual auditing cannot be performed on all work orders, and only a sampling method can be used for auditing, which may miss some cases; manual auditing is also limited by individual ability and division of labor: the same photo may be normal in the scope of the work order of auditor A, but it is a fake photo in the scope of the work order of auditor B, so the number of fake photos identified by manual auditing is much less than the actual number of fake photos.

[0055] Based on this, the embodiments of the present disclosure provide a picture auditing method, device, electronic equipment and storage medium, which can intelligently identify work photos based on a YOLO (You Only Look Once) target detection algorithm, and realize automatic auditing of maintenance work photos (referred to as pictures) by combining a picture similarity algorithm. The picture auditing method is mainly aimed at improving the efficiency and accuracy of maintenance work photo auditing, quickly identifies the key content in each business scenario picture through the YOLO target detection algorithm, and combines and applies algorithms such as mean hash algorithm, difference hash algorithm, perceptual hash algorithm, and scale-invariant feature transform algorithm (SIFT), to realize full-quantity maintenance work photo auditing and improve the accuracy of picture auditing. Integrating an automatic auditing tool in the maintenance management system can timely audit work pictures, reduce the picture auditing processing time, and improve the picture auditing efficiency; at the same time, the present disclosure only needs to spend a small amount of manpower to manually intervene in the pictures that cannot be identified by the automatic auditing tool, greatly reducing the labor cost, and thus achieving the purpose of improving the picture auditing effect and customer satisfaction.

[0056] The picture auditing method, device, electronic equipment and storage medium provided by the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0057] FIG. 1 is a flowchart of a picture auditing method according to an embodiment of the present disclosure. The picture auditing method can be applied to a maintenance management system, which can be installed in a server for executing the picture auditing method. As shown in FIG. 1, the picture auditing method can include the following steps.

[0058] S101, obtaining a picture to be audited.

[0059] The to-be-audited pictures are at least two.

[0060] In the embodiments of the present disclosure, when the picture auditing method is executed, the to-be-audited pictures can be obtained first, for example, two to-be-audited pictures can be obtained. The to-be-audited pictures can include real-time uploaded maintenance auditing pictures, or can include previously uploaded historical maintenance auditing pictures.

[0061] S102, a preset algorithm is used to calculate the similarity between the to-be-audited pictures.

[0062] The preset algorithm includes at least two similarity calculation algorithms.

[0063] In the embodiments of the present disclosure, after the to-be-audited pictures are obtained, the similarity between the two to-be-audited pictures can be calculated by using a preset algorithm. The preset algorithm can be a preset similarity calculation method, for example, an average hash algorithm, a difference hash algorithm, a perceptual hash algorithm, and the like. In this way, by using the preset algorithm to calculate the similarity between the to-be-audited pictures, at least two similarities between the to-be-audited pictures can be obtained.

[0064] S103, whether the to-be-audited pictures are similar pictures is determined according to the plurality of similarities between the to-be-audited pictures.

[0065] In the embodiments of the present disclosure, after the at least two similarities between the to-be-audited pictures are calculated by using the preset algorithm, whether the to-be-audited pictures are similar pictures can be determined according to the plurality of similarities between the to-be-audited pictures. For example, when the similarity between the to-be-audited pictures belongs to a certain range, it can be determined that the to-be-audited pictures are similar pictures. Taking the to-be-audited pictures including a maintenance work picture 1 and a maintenance work picture 2 as an example, the similarity 1 and the similarity 2 between the maintenance work picture 1 and the maintenance work picture 2 can be calculated by using a similarity calculation algorithm 1 and a similarity calculation algorithm 2 respectively. Then, whether the similarity 1 and the similarity 2 satisfy a preset similarity condition is determined, for example, whether one or both of the similarity 1 and the similarity 2 is in a preset similarity range. If the preset similarity condition is satisfied, it can be considered that the to-be-audited pictures are similar pictures; otherwise, it can be considered that the to-be-audited pictures are not similar pictures.

[0066] In the embodiments of the present disclosure, a picture to be audited is obtained; the picture to be audited is at least two; a preset algorithm is used to calculate the similarity between the pictures to be audited; wherein the preset algorithm includes at least two similarity calculation algorithms; according to the plurality of similarities between the pictures to be audited, it is determined whether the pictures to be audited are similar pictures. In this way, the similarity between the pictures to be audited can be calculated by multiple algorithms, and whether the pictures to be audited are similar pictures can be determined by the similarity. In this way, the automatic auditing processing of the maintenance work pictures can be realized by multiple similarity algorithms. Compared with manual auditing processing, not only can the judgment errors easily occurred in manual auditing be effectively avoided, the picture auditing efficiency can be effectively improved, and the accuracy and integrity of the picture auditing result can be improved; but also it can be applied to the case that the number of maintenance work pictures is large, and the applicability of the picture auditing method can be improved.

[0067] In a possible implementation, before the similarity between the pictures to be audited is calculated by using the preset algorithm, the following further includes:

[0068] The preset target information in each picture to be audited is determined by using a preset artificial intelligence (AI) recognition model; wherein the preset target information includes a preset target and the number of each preset target;

[0069] In the case that the preset target information between the pictures to be audited is different, it is determined that the pictures to be audited are not similar pictures;

[0070] The similarity between the pictures to be audited is calculated by using the preset algorithm, including:

[0071] In the case that the preset target information between the pictures to be audited is the same, the similarity between the pictures to be audited is calculated by using the preset algorithm.

[0072] In the embodiments of the present disclosure, the pre-picture information comparison can also be performed on the to-be-audited pictures first, and only when the picture information in the to-be-audited pictures is different, the similarity is calculated to determine whether the to-be-audited pictures are similar pictures. For example, the picture information in each to-be-audited picture, i.e., the preset target information, can be identified by a preset AI recognition model first. The preset target information can include a preset target and a number of preset targets. The preset AI recognition model can be a pre-trained model for identifying the preset target information in the to-be-audited pictures. For example, the preset AI recognition model can be trained by a YOLO deep learning target detection algorithm. The preset target can be determined according to a business scenario and can be a key target, such as a management label, a resource device (for example, a base station, a tower and a feeder, a building base band unit (BBU), a remote radio unit (RRU), etc.), a scene identifier, a maintenance personnel, and the like. For specific examples, refer to Table 1. Different AI intelligent recognition models can be constructed according to a business scenario and a key target list.

[0073] Table 1

[0074] The target detection model, i.e., the preset AI recognition model, can be trained according to the key target types of different business scenarios. The production environment historical maintenance operation photos are classified according to the business scenarios to form a training data set of different preset AI recognition models. Then, the key targets in various maintenance operation photos are labeled, for example, the key target of “maintenance personnel wearing a safety helmet” is labeled. The preset AI recognition model is trained by using the YOLO deep learning target detection algorithm and the labeled maintenance operation photo training set, and the model parameters are adjusted according to the recognition result each time, so that the loss function of the key target type part of the trained model meets the set threshold and the key target type recognition accuracy meets the requirements. After the identification of the preset AI recognition model is completed, the weight parameter file of the preset AI recognition model of different business scenarios is obtained, which supports the identification of key target types such as management labels, resource devices, and emergency vehicles. The to-be-audited pictures, i.e., the to-be-audited maintenance operation photos, are identified, for example, the safety helmet and the maintenance personnel wearing the safety helmet can be extracted, and the identification result can be, for example: 2 maintenance personnel wearing safety helmets are identified.

[0075] After determining the preset targets contained in each of the to-be-audited pictures and the quantity of each preset target by the preset AI recognition model, it can be determined whether the picture information between the two to-be-audited pictures is the same, that is, whether the preset targets contained in the two to-be-audited pictures and the quantity of each preset target are the same. If the preset targets contained in the two to-be-audited pictures and the quantity of each preset target are not the same, then the two to-be-audited pictures are definitely not similar photos. Conversely, if the preset targets contained in the two to-be-audited pictures and the quantity of each preset target are the same, that is, the preset target information between the to-be-audited pictures is the same, then it is necessary to further audit and confirm whether the two pictures are similar pictures by calculating the similarity. In this way, the preset targets and the quantity of each preset target in the work photos can be identified by the AI recognition model first, and it can be quickly confirmed whether the two pictures are similar photos: if the quantity of key targets is inconsistent, then the two pictures are definitely not similar photos; if the quantity of key targets is consistent, then further auditing is needed to confirm whether the two pictures are similar photos.

[0076] It can be understood that after the preset AI recognition model identification is completed, manual sampling inspection can be performed on the identification result, re-labeling learning can be performed based on the pictures that fail to be identified, and the newly added audited and passed pictures can be used as incremental samples to supplement the training data set of the preset AI recognition model. The newly added audited result pictures have a positive feedback effect on the training of the preset AI recognition model, and through dynamic data training, the AI recognition model can be continuously adjusted and optimized to make the identification result more accurate.

[0077] In a possible implementation, the similarity calculation algorithm includes a mean hash algorithm, a difference hash algorithm, and a perceptual hash algorithm.

[0078] The preset algorithm is used to calculate the similarity between the to-be-audited pictures, including:

[0079] The mean hash algorithm, the difference hash algorithm, and the perceptual hash algorithm are used respectively to calculate the first similarity, the second similarity, and the third similarity between the to-be-audited pictures.

[0080] According to the plurality of similarities between the to-be-audited pictures, it is determined whether the to-be-audited pictures are similar pictures, including:

[0081] According to the first similarity, the second similarity, and the third similarity, it is determined whether the to-be-audited pictures are similar pictures.

[0082] In the embodiments of the present disclosure, the similarity calculation algorithm can include a mean hash algorithm, a difference hash algorithm, a perceptual hash algorithm, or can also include other similarity calculation algorithms. When the preset algorithm is used to calculate the similarity between the to-be-audited pictures, the mean hash algorithm can be used to calculate the similarity (i.e., the first similarity) between the to-be-audited pictures, the difference hash algorithm can be used to calculate the similarity (i.e., the second similarity) between the to-be-audited pictures, and the perceptual hash algorithm can be used to calculate the similarity (i.e., the third similarity) between the to-be-audited pictures. Then, whether the to-be-audited picture is a similar picture can be determined according to the calculated first similarity, second similarity, and third similarity. It can be understood that the value range of the similarity can be [0, 1], and the closer to 0, the lower the similarity, and the closer to 1, the higher the similarity.

[0083] For example, the evaluation of the picture similarity calculation algorithm can be performed in advance. For example, the original picture and different pictures, pictures that have been cropped, added filters, original pictures, pictures that have been rotated, and pictures that have added subtitles can be used as algorithm inputs, the similarity between each picture and the original picture can be calculated, and the similarity and performance (e.g., time consumption) can be recorded. Specific examples can be shown in Table 2. As can be seen, for the four pictures of cropping, filter, original picture, and rotation, the greater the similarity, the higher the algorithm accuracy. For different pictures, the smaller the similarity, the higher the algorithm accuracy. For the non-rotation scene, the performance and accuracy of the mean hash algorithm, the difference hash algorithm, and the perceptual hash algorithm are all high, and the accuracy of the SIFT algorithm is high in the rotation scene, but the performance of the SIFT algorithm is poor. Therefore, the mean hash algorithm, the difference hash algorithm, and the perceptual hash algorithm are used preferentially.

[0084] Table 2

[0085] In a further possible implementation, whether the to-be-audited picture is a similar picture is determined according to the first similarity, the second similarity, and the third similarity, including:

[0086] The minimum similarity in the first similarity, the second similarity, and the third similarity is selected.

[0087] In a case where the minimum similarity is less than the first preset similarity threshold, the to-be-audited picture is determined to be a dissimilar picture.

[0088] In the embodiments of the present disclosure, when determining whether the to-be-audited pictures are similar pictures according to the first similarity, the second similarity and the third similarity, the minimum value of the three similarities can be selected first, that is, the minimum similarity of the first similarity, the second similarity and the third similarity is selected. Then, it can be determined whether the minimum similarity is less than a first preset similarity threshold, for example, the first preset similarity threshold can be set to 0.45. If the minimum similarity of the first similarity, the second similarity and the third similarity is less than the first preset similarity threshold, it can be determined that the to-be-audited pictures are not similar pictures. In this way, the performance of the three similarities is higher, the minimum similarity is used to determine the similar pictures, and the performance and accuracy of the picture auditing method can be improved.

[0089] In a further possible implementation, the picture auditing method further includes:

[0090] In the case where the minimum similarity is greater than or equal to the first preset similarity threshold, it is determined whether the minimum similarity is greater than a second preset similarity threshold;

[0091] In the case where the minimum similarity is greater than the second preset similarity threshold, the to-be-audited pictures are determined to be similar pictures.

[0092] In the embodiments of the present disclosure, if the minimum similarity of the first similarity, the second similarity and the third similarity is greater than or equal to the first preset similarity threshold, the second preset similarity threshold can be obtained, for example, the second preset similarity threshold can be set to 0.75. The minimum similarity of the first similarity, the second similarity and the third similarity is compared with the second preset similarity threshold to determine whether the minimum similarity is greater than the second preset similarity threshold. If the minimum similarity is greater than the second preset similarity threshold, the to-be-audited pictures are determined to be similar pictures. In this way, only in the case where the minimum similarity of the three similarities is greater than the second preset similarity threshold, the to-be-audited pictures are determined to be similar pictures, and thus the accuracy of the auditing result can be further improved.

[0093] In a further possible implementation, the similarity calculation algorithm further includes a scale-invariant feature transform algorithm. The picture auditing method further includes:

[0094] In the case where the minimum similarity is less than the second preset similarity threshold, a fourth similarity between the to-be-audited pictures is calculated by using the scale-invariant feature transform algorithm;

[0095] In the case where the fourth similarity is greater than the second preset similarity threshold, the to-be-audited pictures are determined to be similar pictures;

[0096] In the case where the fourth similarity is less than or equal to the second preset similarity threshold, the to-be-audited pictures are determined to be not similar pictures.

[0097] In the embodiments of the present disclosure, if the minimum similarity is less than the second preset similarity threshold, the scale-invariant feature transform (SIFT) is used to calculate the similarity between the two pictures to be audited again, that is, the fourth similarity. Then the fourth similarity is compared with the second preset similarity threshold to determine whether the fourth similarity is greater than the second preset similarity threshold. If the fourth similarity is greater than the second preset similarity threshold, it is determined that the two pictures to be audited are similar pictures. Otherwise, if the fourth similarity is less than or equal to the second preset similarity threshold, it is determined that the two pictures to be audited are dissimilar pictures. In this way, on the basis of the mean hash algorithm, the difference hash algorithm, and the perceptual hash algorithm, the SIFT algorithm is further combined to implement picture auditing, so that the picture auditing processing time can be further reduced, and the picture auditing efficiency and accuracy can be improved.

[0098] To make the method provided by the embodiments of the present disclosure clearer, the picture auditing method provided by the embodiments of the present disclosure is described below in combination with FIG. 2. As shown in FIG. 2, the picture auditing method can include the following processing:

[0099] First, according to the work order business scenario, a preset AI recognition model is confirmed to identify and calculate the preset target and the number of preset targets in the two pictures to be audited. If the preset target or the number of preset targets in the two pictures to be audited is inconsistent, it is determined that the two pictures to be audited are dissimilar pictures. If the preset target and the number of preset targets in the two pictures to be audited are consistent, the picture similarity is calculated using a preset algorithm (hash algorithm). The similarity value range is [0, 1], the closer to 0, the lower the similarity, and the closer to 1, the higher the similarity. For the input two pictures to be audited, the similarity is calculated using three high-performance hash algorithms (mean hash algorithm, difference hash algorithm, and perceptual hash algorithm), and the minimum similarity calculated by the three algorithms is taken as the comparison result. If the minimum similarity is low (for example, less than the first preset similarity threshold 0.45), it is determined that the two pictures to be audited are dissimilar pictures, and the calculation is ended. If the minimum similarity is high (for example, greater than the second preset similarity threshold 0.75), it is determined that the two pictures to be audited are similar pictures, and the calculation is ended. If the minimum similarity is less than 0.75, the SIFT algorithm is used to calculate the fourth similarity of the two pictures to be audited. If the fourth similarity is high (for example, greater than the second preset similarity threshold 0.75), it is determined that the two pictures to be audited are similar pictures, and the calculation is ended. Otherwise, it is determined that the two pictures to be audited are dissimilar pictures, and the calculation is ended. In this way, the performance and accuracy of the scheme are taken into account.

[0100] Based on this, the picture auditing method provided by the embodiments of the present disclosure can be based on an AI intelligent auditing method and system for maintenance work photos. According to the work order business scenario, a preset AI recognition model is constructed to calculate the preset target and quantity of the pictures to be audited, judge the similarity of the pictures to be audited, and analyze and evaluate from multiple scenes such as cropping, filters, rotation, and watermarks through multiple hash algorithms combined with algorithms such as scale-invariant feature transform, and then judge the picture similarity to realize automatic auditing of maintenance work photos. In this way, the automatic auditing tool replaces manual auditing, which can greatly reduce the long time consumption of manual auditing and the large consumption of human resources, solve the problem of large workload of manual auditing and the problem of missed judgment of cross-maintenance work order auditing, reduce the risk of maintenance management, and improve the level of maintenance management, thereby reducing the processing time of auditing, improving the efficiency of auditing, and reducing the cost of human resources. Moreover, the auditing personnel do not need to actively audit, but only need to confirm a small amount of suspicious work order sign-in photos to complete the auditing of all work order sign-in photos, fully improve the integrity of the auditing, more effectively standardize the work of the maintenance personnel, and ultimately achieve the purpose of improving customer satisfaction.

[0101] For example, the execution system of the picture auditing method can include an interface layer, a business layer, and an algorithm layer, as shown in FIG. 3. The business layer calls the algorithm layer to realize the degree of similarity calculation, and provides the business capability to the interface layer. As an example, the system realizes performance optimization by limiting the range of compared pictures. Since similar photos are generally generated only from work orders of the same business scenario within the same maintenance team, we limit the range of two-by-two comparison, rather than calculating 100000*100000 times of similarity. For one picture, we calculate according to the range of maintenance team + work order type + business scenario + archived work orders within 7 days. The maintenance team is a team responsible for the maintenance work of a region, generally composed of about 5 professional maintenance personnel. The work order type is a large category of business managed by the work order, such as wireless work order, transmission work order, etc. The business scenario is a specific work classification in the large category of business, as shown in Table 1. The pictures submitted by the work orders of the same business scenario of the same maintenance team are generally not more than 20 per day, the maximum calculation amount is 140*100000, and the average calculation time is not more than 1 second. Using 5 32-core servers can realize the auditing of all work orders.

[0102] Based on the same inventive concept, the embodiments of the present disclosure also provide a picture auditing device, as shown in FIG. 4, which is a block diagram of a picture auditing device according to an embodiment of the present disclosure. Referring to FIG. 4, the picture auditing device 400 can include:

[0103] The picture receiving module 410 is configured to receive pictures to be audited; the pictures to be audited are at least two;

[0104] The calculation module 420 is configured to calculate the similarity between the to-be-audited pictures by using a preset algorithm; the preset algorithm includes at least two similarity calculation algorithms.

[0105] The auditing module 430 is configured to determine whether the to-be-audited pictures are similar pictures according to the similarities between the to-be-audited pictures.

[0106] In a possible implementation, the picture auditing apparatus 400 further includes:

[0107] The target identification module is configured to determine preset target information in each to-be-audited picture by using a preset AI identification model; the preset target information includes preset targets and the number of each preset target.

[0108] The first determination module is configured to determine that the to-be-audited pictures are dissimilar pictures when the preset target information between the to-be-audited pictures is different.

[0109] The calculation module 420 is configured to:

[0110] When the preset target information between the to-be-audited pictures is the same, the calculation module 420 is configured to calculate the similarity between the to-be-audited pictures by using a preset algorithm.

[0111] In a possible implementation, the similarity calculation algorithm includes a mean hash algorithm, a difference hash algorithm, and a perceptual hash algorithm.

[0112] The calculation module 420 is configured to:

[0113] The calculation module 420 is configured to calculate a first similarity, a second similarity, and a third similarity between the to-be-audited pictures by using the mean hash algorithm, the difference hash algorithm, and the perceptual hash algorithm, respectively.

[0114] The auditing module 430 is configured to:

[0115] The auditing module 430 is configured to determine whether the to-be-audited pictures are similar pictures according to the first similarity, the second similarity, and the third similarity.

[0116] In a possible implementation, the auditing module 430 includes:

[0117] The selecting unit is configured to select a minimum similarity from the first similarity, the second similarity, and the third similarity.

[0118] The auditing unit is configured to determine that the to-be-audited pictures are dissimilar pictures when the minimum similarity is less than a first preset similarity threshold.

[0119] In a possible implementation, the picture auditing apparatus 400 further includes:

[0120] The similarity judgment module is configured to determine whether the minimum similarity is greater than a second preset similarity threshold value in a case where the minimum similarity is greater than or equal to the first preset similarity threshold value.

[0121] The second determination module is configured to determine that the picture to be audited is a similar picture in a case where the minimum similarity is greater than the second preset similarity threshold value.

[0122] In a possible implementation, the similarity calculation algorithm further includes a scale-invariant feature transform algorithm.

[0123] The picture auditing apparatus 400 further includes:

[0124] The similarity calculation module is configured to calculate a fourth similarity between the pictures to be audited by using the scale-invariant feature transform algorithm in a case where the minimum similarity is less than the second preset similarity threshold value.

[0125] The third determination module is configured to determine that the picture to be audited is a similar picture in a case where the fourth similarity is greater than the second preset similarity threshold value.

[0126] The fourth determination module is configured to determine that the picture to be audited is a dissimilar picture in a case where the fourth similarity is less than or equal to the second preset similarity threshold value.

[0127] As to the apparatus in the above-described embodiments, specific manners in which various modules perform operations have been described in details in the embodiments of the method, and will not be described here in details.

[0128] According to embodiments of the present disclosure, the present disclosure further provides an electronic device, a storage medium, and a computer program product.

[0129] FIG. 5 shows a schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present disclosure. The electronic device 500 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present disclosure described and / or claimed in this document.

[0130] As shown in FIG. 5, the electronic device 500 includes a computing unit 501 that can perform various appropriate actions and processes in accordance with a computer program stored in a Read-Only Memory (ROM) 502 or a computer program loaded into a Random Access Memory (RAM) 503 from a storage unit 508. Various programs and data required for the operation of the electronic device 500 can also be stored in the RAM 503. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An Input / Output (I / O) interface 505 is also connected to the bus 504.

[0131] Various components in the electronic device 500 are connected to the I / O interface 505, including an input unit 506 such as a keyboard, a mouse, and the like, an output unit 507 such as various types of displays, a speaker, and the like, a storage unit 508 such as a magnetic disk, an optical disk, and the like, and a communication unit 509 such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0132] The computing unit 501 can be various general-purpose and / or special-purpose processing components having processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processing (DSP), and any appropriate processor, controller, microcontroller, and the like. The computing unit 501 performs various methods and processes described above, such as the picture auditing method. For example, in some embodiments, the picture auditing method can be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the picture auditing method described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the picture auditing method by any other appropriate means, such as by means of firmware.

[0133] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an application specific standard product (ASSP), a system on chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0134] Program code of a computer program product for implementing the methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, implements the functions / operations specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as part of a separate software package, and partially on a remote machine or server.

[0135] In the context of this disclosure, a storage medium can be a tangible medium which can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A storage medium can be a machine-readable signal medium or a machine-readable storage medium. Storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of storage media can include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fiber, compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0136] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a Cathode Ray Tube (CRT) or Liquid Crystal Display (LCD) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, 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, speech, or tactile input.

[0137] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can 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), the Internet, and a blockchain network.

[0138] The computer system can include clients and servers. This relationship can be implemented by a computer program running on the respective computers and having a client-server relationship with one another. The servers can be cloud servers, also known as cloud computing servers or cloud hosts, which are a host product in the cloud computing service system to solve the defects of large management difficulty and weak business scalability in traditional physical hosts and virtual private servers (VPS). The servers can also be servers of a distributed system or servers combined with a blockchain.

[0139] It should be understood that the steps shown above can be reordered, added to, or deleted from. For example, the steps described in the present disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions of the present disclosure can be achieved, and the present disclosure is not limited herein.

[0140] The above detailed description does not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A picture auditing method, comprising: obtaining pictures to be audited; the pictures to be audited are at least two; calculating the similarity between the pictures to be audited by using a preset algorithm; wherein the preset algorithm comprises at least two similarity calculation algorithms; determining whether the pictures to be audited are similar pictures according to a plurality of similarities between the pictures to be audited.

2. The picture vetting method of claim 1, wherein, Before the similarity between the pictures to be audited is calculated by using the preset algorithm, further comprising: determining a preset target information in each of the pictures to be audited by a preset artificial intelligence (AI) recognition model; wherein the preset target information comprises a preset target and the number of each preset target; in the case that the preset target information between the pictures to be audited is different, determining that the pictures to be audited are not similar pictures; the similarity between the pictures to be audited is calculated by using a preset algorithm, comprising: in the case that the preset target information between the pictures to be audited is the same, calculating the similarity between the pictures to be audited by using a preset algorithm.

3. The picture vetting method of claim 1 or 2, wherein, The similarity calculation algorithm comprises a mean hash algorithm, a difference hash algorithm, and a perceptual hash algorithm; the similarity between the pictures to be audited is calculated by using a preset algorithm, comprising: respectively using the mean hash algorithm, the difference hash algorithm, and the perceptual hash algorithm to calculate a first similarity, a second similarity, and a third similarity between the pictures to be audited; determining whether the pictures to be audited are similar pictures according to a plurality of similarities between the pictures to be audited, comprising: determining whether the pictures to be audited are similar pictures according to the first similarity, the second similarity, and the third similarity.

4. The picture bugging method of claim 3, wherein, determining whether the pictures to be audited are similar pictures according to the first similarity, the second similarity, and the third similarity, comprising: selecting the minimum similarity among the first similarity, the second similarity, and the third similarity; in the case that the minimum similarity is less than a first preset similarity threshold, determining that the pictures to be audited are not similar pictures.

5. The picture auditing method of claim 4, further comprising: in the case that the minimum similarity is greater than or equal to the first preset similarity threshold, determining whether the minimum similarity is greater than a second preset similarity threshold; in the case that the minimum similarity is greater than the second preset similarity threshold, determining that the pictures to be audited are similar pictures.

6. The picture bugging method of claim 5, wherein, The similarity calculation algorithm further comprises a scale-invariant feature transform algorithm; The picture auditing method further comprises: in the case that the minimum similarity is less than the second preset similarity threshold, calculating a fourth similarity between the pictures to be audited by using the scale-invariant feature transform algorithm; in the case that the fourth similarity is greater than the second preset similarity threshold, determining that the pictures to be audited are similar pictures; in the case that the fourth similarity is less than or equal to the second preset similarity threshold, determining that the pictures to be audited are not similar pictures.

7. A picture auditing device, comprising: a picture receiving module for receiving pictures to be audited; the pictures to be audited are at least two; a calculation module, configured to calculate similarity between the to-be-audited pictures by using a preset algorithm; wherein the preset algorithm comprises at least two similarity calculation algorithms; an auditing module, configured to determine whether the to-be-audited pictures are similar pictures according to the similarities between the to-be-audited pictures. 8.An electronic device, comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the picture auditing method according to any one of claims 1 to 6. 9.A storage medium, when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the picture auditing method according to any one of claims 1 to 6. 10.A computer program product, comprising a computer program, when the computer program is executed by a processor, the picture auditing method according to any one of claims 1 to 6 is implemented.

Images