Method, device, equipment and storage medium for corrosion detection of a tank

Abstract

The application discloses a rust detection method, device and equipment for a storage tank and a storage medium, and the method comprises the following steps: establishing a tank wall three-dimensional image of a detected storage tank according to detection data comprising three-dimensional point cloud data of a tank wall of the storage tank, and generating a first side development drawing of a tank wall surface of the detected storage tank according to the tank wall three-dimensional image through picture splicing; performing rust mark identification on the first side development drawing through a tank wall rust mark identification model based on an artificial neural network, and determining rust pixels from the first side development drawing; calculating a tank wall rust area ratio of the tank wall of the storage tank; and generating a tank wall detection result of the tank wall of the storage tank according to the tank wall rust area ratio. The application excludes a human subjective judgment link, and therefore, the determination result is more accurate and stable.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of chemical equipment, in particular to a rust detection method, device, equipment and storage medium for a storage tank. BACKGROUND

[0002] The storage tank is a large container for storing liquid medium, which is commonly used in oil refineries, oil fields, oil depots and other refining scenes. Tanks form a tank farm to store various media. The storage tank is widely used for storing liquid medium in oil fields, refineries, stations and depots. During operation, the tank body is often corroded due to external environmental factors.

[0003] According to the storage tank maintenance and repair regulations, the corrosion degree of the storage tank needs to be periodically detected to make maintenance and repair decisions.

[0004] In the prior art, the determination of the rust condition of the tank body generally needs to be compared by manual naked eye.

[0005] The inventors have found that the existing method for determining the rust condition of the tank body has at least the following defects:

[0006] Since the manual comparison method lacks objective quantitative indicators and does not have a unified standard, the determination of the rust condition of the tank body by manual method is prone to poor stability and accuracy of the determination results due to personal subjective factors. SUMMARY

[0007] The main purpose of the present application is to improve the stability and accuracy of the rust detection for the storage tank.

[0008] To achieve the above purpose, the technical solution adopted by the present application is:

[0009] The present application discloses a rust detection method for a storage tank, comprising:

[0010] S11, establishing a tank wall three-dimensional image of the detected storage tank according to the detection data including three-dimensional point cloud data of the tank wall of the storage tank, and generating a first side development drawing of the tank wall surface of the detected storage tank according to the tank wall three-dimensional image through picture stitching; the three-dimensional point cloud data is obtained by a three-dimensional laser scanner;

[0011] S12, rust mark recognition is performed on the first side development drawing by a tank wall rust mark recognition model based on artificial neural network, and rust pixels are determined from the first side development drawing; the tank wall rust mark recognition model is constructed by taking the first side development drawing of the tank wall of the storage tank generated in the past as modeling data and through data training, and is used for recognizing rust marks on the tank wall of the storage tank;

[0012] S13, calculate a tank wall corrosion area ratio of the tank wall by counting the corrosion pixels of the first side unfolding map; the tank wall corrosion area ratio is a ratio of a corrosion area to an area of the first side unfolding map;

[0013] S14, generate a tank wall detection result of the tank wall according to the tank wall corrosion area ratio.

[0014] Preferably, in the present application, further comprising:

[0015] S21, the detection data further comprises tank wall scanning data, a three-dimensional SAR image of the tank wall of the detected tank is established according to the tank wall scanning data, and a second side unfolding map of the tank wall surface of the detected tank is obtained by picture stitching according to the three-dimensional SAR image of the tank wall; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner scanning the tank wall;

[0016] S22, convert the area size and coordinate system of the second side unfolding map into the area size and coordinate system of the first side unfolding map; Figure One

[0017] S23, according to the corrosion pixels in the first side unfolding map, determine the corresponding tank wall reflection rust mark area in the second side unfolding map;

[0018] S24, determine the corrosion depth of the tank wall reflection rust mark area according to the gray value of the tank wall reflection rust mark area.

[0019] Preferably, in the present application, the determination of the corrosion depth of the tank wall reflection rust mark area according to the gray value of the tank wall reflection rust mark area comprises:

[0020] a preset corresponding relationship between the gray value of the SAR image and the corrosion depth;

[0021] determine the corrosion depth of the tank wall reflection rust mark area according to the gray value of the tank wall reflection rust mark area and the corresponding relationship.

[0022] Preferably, in the present application, further comprising:

[0023] According to the corrosion depth of the tank wall reflection rust mark area, divide the tank wall reflection rust mark area into different corrosion degrees.

[0024] Preferably, in the present application, the division of the tank wall reflection rust mark area into different corrosion degrees according to the corrosion depth of the tank wall reflection rust mark area comprises:

[0025] a preset corresponding relationship between different corrosion depth ranges and corrosion degrees;

[0026] ​According to the correspondence, the area or proportion of each rust degree in the rust mark area reflected by the tank wall is determined.

[0027] Preferably, in the present application, further comprising:

[0028] S31, the detection data further comprises a tank top detection photo; a tank top image of the detected storage tank is generated according to the tank top detection photo; the tank top detection photo is obtained by an image acquisition device arranged on the unmanned aerial vehicle; the tank top detection photo is an orthographic image;

[0029] S32, rust mark recognition is performed on the tank top image by a tank top rust mark recognition model based on artificial neural network, and rust pixels are determined from the tank top image; the tank top rust mark recognition model is constructed by taking the tank top images of the tank top of the storage tank generated in the past as modeling data and through data training, and is used for recognizing rust marks on the tank top of the storage tank;

[0030] S33, the tank top rust area ratio of the tank top of the storage tank is calculated by counting the rust pixels of the tank top image;

[0031] S34, a tank top detection result of the tank top of the storage tank is generated according to the tank top rust area ratio.

[0032] Preferably, in the present application, further comprising:

[0033] S41, the detection data further comprises tank top scanning data; a tank top SAR image consistent with the tank top detection photo coordinate system corresponding to the tank top scanning data is generated according to the tank top scanning data; the unmanned aerial vehicle is further provided with a synthetic aperture radar for obtaining tank top scanning data synchronized with the image acquisition device; the tank top scanning data is synchronized and one-to-one corresponding with the tank top detection photo obtained by the image acquisition device;

[0034] S42, according to the rust pixels of the tank top image, a corresponding tank top reflected rust mark area is determined in the tank top SAR image;

[0035] S43, according to the gray value of the tank top reflected rust mark area, the rust depth of the tank top reflected rust mark area is determined.

[0036] Preferably, in the present application, the tank top rust mark recognition model comprises:

[0037] A rectangular recognition sub-model is used to generate a rectangular frame label of the tank top of the detected storage tank according to the tank top detection photo; the rectangular recognition sub-model is generated by training the original tank top detection photo combined with the rectangular frame label; the rectangular frame label of the tank top pattern is obtained by frame selection on the original tank top detection photo as historical data, and a tank top mask label and a rust mark label are generated;

[0038] a mask recognition sub-model, configured to generate a tank top mask label of the to-be-detected storage tank in the tank top rectangular image according to the tank top rectangular image; the mask recognition sub-model is generated by training the tank top rectangular image combined with the tank top mask label;

[0039] a preliminary rust identification sub-model, configured to obtain a preliminary rust identification result according to the tank top rectangular image; the preliminary rust identification sub-model is generated by training the tank top rectangular image combined with the rust label;

[0040] a tank top identification result module, configured to determine rust pixels from the tank top image according to the tank top mask label generated by the mask recognition sub-model and the preliminary rust identification result.

[0041] In another aspect of the present application, a rust detection device for a storage tank is also provided, comprising:

[0042] a tank wall picture generation unit, configured to establish a tank wall three-dimensional image of a to-be-detected storage tank according to detection data including three-dimensional point cloud data of a tank wall of the storage tank, and generate a first side development drawing of a tank wall surface of the to-be-detected storage tank according to the tank wall three-dimensional image by picture stitching; the three-dimensional point cloud data is obtained by a three-dimensional laser scanner;

[0043] a tank wall rust identification unit, configured to determine rust pixels from the first side development drawing by a tank wall rust identification model based on an artificial neural network; the tank wall rust identification model is constructed by data training with modeling data of previously generated first side development drawings of storage tank walls, and is used to identify rust on the tank wall of the storage tank;

[0044] a tank wall rust area ratio calculation unit, configured to calculate a tank wall rust area ratio of the tank wall of the storage tank by counting the rust pixels of the first side development drawing; the tank wall rust area ratio is a ratio of a rust area to an area of the first side development drawing;

[0045] a tank wall detection result generation unit, configured to generate a tank wall detection result of the tank wall of the storage tank according to the tank wall rust area ratio.

[0046] Preferably, in the present application, further comprising:

[0047] a tank wall SAR image generation unit, the detection data further comprises tank wall scanning data, a three-dimensional SAR image of a tank wall of a detected storage tank is established according to the tank wall scanning data, and a second side development drawing of a tank wall surface of the detected storage tank is obtained according to the three-dimensional SAR image of the tank wall by picture splicing; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner and scanning the tank wall of the storage tank;

[0048] a tank wall SAR image adjustment unit, the area size and coordinate system of the second side development drawing are converted into the first side development drawing Figure One ;

[0049] a tank wall rust area determination unit, for determining a corresponding tank wall reflection rust area in the second side development drawing according to rust pixels in the first side development drawing;

[0050] a tank wall rust depth determination unit, for determining a rust depth of the tank wall reflection rust area according to a gray value of the tank wall reflection rust area.

[0051] In another aspect of the embodiment of the present application, a rust detection device for a storage tank is further provided, comprising:

[0052] a memory, for storing a computer program;

[0053] a processor, for calling and executing the computer program to realize each step of the rust detection method for a storage tank according to any one of the above.

[0054] In another aspect of the embodiment of the present application, a storage medium having a computer program stored thereon is further provided, the computer program is executed by a processor to realize each step of the rust detection method for a storage tank according to any one of the above.

[0055] Advantages

[0056] The present application firstly utilizes the imaging technology of the three-dimensional laser scanner to generate a side wall development drawing of the tank wall of the detected storage tank; then, the rust marks of the tank wall of the detected storage tank are identified and calculated through the side wall development drawing by utilizing the image recognition technology based on artificial neural network, and accurate rust mark data is obtained; since the tank wall detection result generated by the embodiment of the present application can accurately and objectively represent the rust degree of the tank wall of the detected storage tank, the stability and accuracy of the rust detection of the storage tank are improved.

[0057] Further, the application can also include tank wall scanning data collected by a synthetic aperture radar synchronized with the three-dimensional laser scanner, and generate a tank wall SAR image consistent with the corresponding first side view development coordinate system according to the tank wall scanning data; since the scattering characteristics of the tank wall SAR image can reflect the roughness of the surface of the photographed object, the corresponding rust area can be determined in the tank wall SAR image according to the rust identification result of the first side view development; then, the corresponding rust depth is determined according to the different scattering characteristics in the tank wall rust area; since through the embodiment of the application, not only the rust area (or rust area ratio) of the tank wall can be determined, but also different rust depths in the rust area can be identified, so that more accurate and comprehensive detection results for the rust degree and rust condition can be obtained.

[0058] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application and can be implemented according to the content of the specification, at the same time, in order to make the above and other purposes, technical features and advantages of the present application more easily understood, one or more preferred embodiments are listed below, and are described in detail as follows with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiment or the prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0060] Figure 1 A schematic diagram of the steps of the corrosion detection method for the storage tank described in the present application;

[0061] Figure 2 A schematic diagram of another step of the corrosion detection method for the storage tank described in the present application;

[0062] Figure 3 A structural schematic diagram of the corrosion detection device for the storage tank described in the present application;

[0063] Figure 4 Another structural schematic diagram of the corrosion detection device for the storage tank described in the present application;

[0064] Figure 5 A structural schematic diagram of the corrosion detection device for the storage tank described in the present application. DETAILED DESCRIPTION

[0065] In order to make the technical personnel in the art better understand the present application scheme, the technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by the technical personnel in the art without creative labor are within the scope of protection of the present application.

[0066] Embodiment one

[0067] In order to improve the stability and accuracy of the corrosion detection of the storage tank, with reference to Figure 1 The present application provides a kind of corrosion detection method for storage tank, comprising the steps of:

[0068] S11, according to the detection data including the three-dimensional point cloud data of the tank wall of the storage tank, the three-dimensional image of the tank wall of the detected storage tank is established, and the first side development drawing of the tank wall surface of the detected storage tank is generated according to the three-dimensional image of the tank wall by picture stitching;The three-dimensional point cloud data is obtained by three-dimensional laser scanner;

[0069] In order to detect the corrosion of the storage tank, in the embodiments of the present application, three-dimensional laser scanner and computer equipment (such as remote processing center) need to be combined to realize, wherein three-dimensional laser scanner is used to obtain three-dimensional point cloud data of the detected storage tank;Computer equipment is provided with tank wall rust mark recognition model based on artificial neural network to identify rust mark.

[0070] In practical application, a plurality of three-dimensional laser scanners can be arranged around the detected storage tank to obtain three-dimensional point cloud data of each view angle of the detected storage tank, so that the computer equipment can construct the three-dimensional image of the tank wall of the detected storage tank (i.e. three-dimensional model of the tank wall).

[0071] The three-dimensional image of the tank wall of the detected storage tank can be developed to obtain a side development drawing of the tank wall (i.e. first side development drawing).

[0072] S12, rust mark recognition model based on artificial neural network is used to identify the first side development drawing, and rust pixel is determined from the first side development drawing;The tank wall rust mark recognition model is constructed by data training through the first side development drawing of the tank wall of the storage tank generated in the past as modeling data, and is used to identify the rust mark on the tank wall of the storage tank;

[0073] In the embodiment of the present application, the tank wall rust mark recognition model based on an artificial neural network can be generated by taking historical data (i.e., a side expansion diagram obtained by expanding a three-dimensional image of a tank wall in the past) as modeling data and through data training. The rust mark on the tank wall of the storage tank is recognized by identifying whether each pixel in the first side expansion diagram is a rust mark pixel.

[0074] S13, calculate a tank wall rust area ratio of the storage tank by counting the rust pixels of the first side expansion diagram; the tank wall rust area ratio is a ratio of a rust area to an area of the first side expansion diagram;

[0075] The tank wall rust area ratio (rust area ratio: a ratio of a rust area on the tank wall to the overall area of the tank wall) of the tank wall of the storage tank is an important basis for judging the rust degree, so in the embodiment of the present application, after the rust marks on the tank wall of the storage tank are recognized, the tank wall rust area ratio of the storage tank is calculated by counting the rust pixels in the first side expansion diagram.

[0076] S14, generate a tank wall detection result of the tank wall of the storage tank according to the tank wall rust area ratio.

[0077] In actual application, a corresponding relationship between a rust area ratio and a rust degree can be determined to generate the corresponding tank wall detection result of the detected storage tank according to the tank wall rust area ratio. Further, a corresponding maintenance strategy can be formulated according to the tank wall detection result.

[0078] In summary, in the embodiment of the present application, the imaging technology of the three-dimensional laser scanner is first used to generate the side wall expansion diagram of the tank wall of the detected storage tank; then, the image recognition technology based on an artificial neural network is used to recognize and calculate the rust marks on the tank wall of the detected storage tank through the side expansion diagram, and accurate rust mark data is obtained. Since the tank wall detection result generated by the embodiment of the present application can accurately and objectively represent the rust degree of the tank wall of the detected storage tank, the stability and accuracy of the rust detection of the storage tank are improved.

[0079] Embodiment Two

[0080] Based on the embodiment one, as shown in the embodiment two, Figure 2 the embodiment of the present application can further include the following steps:

[0081] S21, the detection data further includes tank wall scanning data, a three-dimensional SAR image of the tank wall of the detected storage tank is established according to the tank wall scanning data, and a second side expansion diagram of the tank wall surface of the detected storage tank is obtained from the three-dimensional SAR image of the tank wall according to picture stitching; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner scanning the tank wall of the storage tank;

[0082] Although the image obtained by the three-dimensional laser scanner (the first side view drawing) can accurately identify the rust on the tank wall through the tank wall rust identification model, the rust depth of the tank wall cannot be identified. Therefore, in the embodiment of the present application, the correlation between the scattering characteristics of the SAR image of the synthetic aperture radar and the roughness of the surface of the scanned object is further utilized to determine the corrosion depth of the surface of the storage tank. Specifically, the SAR image of the scanned object obtained by the synthetic aperture radar can be used to infer the roughness of the surface of the scanned object according to the scattering characteristics (such as the gray value) of the SAR image. For the storage tank, the higher the roughness of the rust on the surface, the deeper the corrosion depth. The corrosion depth is also one of the important indicators for determining the corrosion degree of the surface of the storage tank.

[0083] In order to maintain consistency with the image obtained by the three-dimensional laser scanner, in the embodiment of the present application, the synthetic aperture radar and the three-dimensional laser scanner need to be arranged at the same position to scan the tank wall of the storage tank at the same angle of view.

[0084] S22, the area size and coordinate system of the second side view drawing are converted to be consistent with the first side view drawing. Figure One ;

[0085] Since the synthetic aperture radar and the three-dimensional laser scanner in the embodiment of the present application scan the tank wall of the storage tank at the same angle of view, the tank wall scanning data and the three-dimensional point cloud data of the tank wall are consistent. Therefore, the second side view drawing generated according to the tank wall scanning data can be converted in the coordinate system with the first side view drawing generated according to the three-dimensional point cloud data of the tank wall as the standard, so that the area size and the coordinate system are consistent with the first side view drawing.

[0086] S23, according to the rust pixels in the first side view drawing, the corresponding tank wall reflection rust area in the second side view drawing is determined.

[0087] Since the rust on the tank wall has been identified by the tank wall rust identification model, the corresponding tank wall reflection rust area can be determined from the second side view drawing according to the identification result.

[0088] S24, according to the gray value of the tank wall reflection rust area, the corrosion depth of the tank wall reflection rust area is determined.

[0089] The scattering characteristic (gray value) of the scanned object surface can be obtained by the synthetic aperture radar, and the typical ones include that the roughness of the scanned object surface determines the gray value (that is, the echo intensity) of the SAR image; the inventor finds that the roughness of the rust stain on the surface of the storage tank is strongly correlated with the corrosion depth, and therefore, in the embodiment of the present application, the scattering characteristic represented by the SAR image is used as the measurement parameter of the corrosion depth, that is, the greater the gray value of the SAR image, the deeper the corrosion depth, and vice versa.

[0090] Specifically, when determining the corrosion depth of the rust stain area reflected by the tank wall, a corresponding relationship between the gray value of the SAR image and the corrosion depth can be preset; and then the corrosion depth of the rust stain area reflected by the tank wall is determined according to the gray value of the rust stain area reflected by the tank wall in the SAR image and the corresponding relationship.

[0091] In actual application, the rust stain area reflected by the tank wall can also be divided into different corrosion degrees according to the corrosion depth of the rust stain area reflected by the tank wall; specifically, it can include:

[0092] a corresponding relationship between different corrosion depth ranges and corrosion degrees is preset;

[0093] the area or proportion of each corrosion degree in the rust stain area reflected by the tank wall is determined according to the corresponding relationship.

[0094] As described above, on the basis of the first embodiment, the embodiment of the present application can further include the tank wall scanning data collected by the synthetic aperture radar synchronized with the three-dimensional laser scanner, and generate the tank wall SAR image consistent with the corresponding first side view development map coordinate system according to the tank wall scanning data; since the scattering characteristic of the tank wall SAR image can reflect the roughness of the surface of the photographed object, the corresponding rust stain area can be determined in the tank wall SAR image according to the rust stain recognition result of the first side view development map; then, the corresponding corrosion depth is determined according to the different scattering characteristics in the tank wall rust stain area; since the embodiment of the present application can not only determine the corrosion area (or the corrosion area ratio) of the tank wall of the storage tank, but also identify the different corrosion depths in the rust stain area, a more accurate and comprehensive detection result for the corrosion degree and the corrosion condition can be obtained.

[0095] Embodiment three

[0096] On the basis of the first embodiment or the second embodiment, the embodiment of the present application further adds the technical scheme of rust stain recognition for the tank top of the detected storage tank, and specifically includes:

[0097] S31, the detection data further includes a tank top detection photo; the tank top image of the detected storage tank is generated according to the tank top detection photo; the tank top detection photo is obtained by the image acquisition device arranged on the unmanned aerial vehicle; the tank top detection photo is an orthographic image;

[0098] In the embodiment of the present application, the unmanned aerial vehicle is also required to collect the tank top detection photos of the tank area; and the computer device can also be used to perform corresponding processing and operation according to the tank top detection photos to obtain the corrosion detection result of the storage tank. Similarly, in the embodiment of the present application, the computer device can also be provided with a tank top rust mark recognition model based on artificial neural network to perform rust mark recognition on the tank top image.

[0099] S32, rust mark recognition is performed on the tank top image by the tank top rust mark recognition model based on artificial neural network to determine rust pixels from the tank top image; the tank top rust mark recognition model is constructed by taking the tank top images of the tank top of the storage tank generated in the past as modeling data and through data training, and is used to recognize rust marks on the tank top of the storage tank;

[0100] In actual application, the tank top rust mark recognition model can specifically include:

[0101] a rectangular recognition sub-model, used to generate a rectangular frame label of the tank top of the detected storage tank according to the tank top detection photo; the rectangular recognition sub-model is generated by training the original tank top detection photo combined with the rectangular frame label; the rectangular frame label of the tank top pattern is obtained by frame selection on the original tank top detection photo as historical data, and a tank top mask label and a rust mark label are generated;

[0102] a mask recognition sub-model, used to generate a tank top mask label of the detected storage tank in the tank top rectangular image according to the tank top rectangular image; the mask recognition sub-model is generated by training the tank top rectangular image combined with the tank top mask label; the tank top rectangular image is obtained according to the rectangular label;

[0103] a preliminary rust mark recognition sub-model, used to obtain a preliminary rust mark recognition result according to the tank top rectangular image; the preliminary rust mark recognition sub-model is generated by training the tank top rectangular image combined with the rust mark label;

[0104] a tank top recognition result module, used to determine rust pixels from the tank top image according to the tank top mask label generated by the mask recognition sub-model and the preliminary rust mark recognition result.

[0105] The tank top rust mark recognition model in the embodiment of the present application includes multiple sub-models (i.e., the rectangular recognition sub-model, the mask recognition sub-model, and the preliminary rust mark recognition sub-model) to respectively complete different sub-functions, wherein the rectangular recognition sub-model is used to generate a rectangular frame label of the detected storage tank in the tank top detection photo according to the tank top detection photo.

[0106] In practical applications, the rectangular recognition sub-model can be trained and generated by combining the original tank top detection photos with the rectangular frame labels; that is, the original tank top detection photos as historical data are used as modeling data for training, which can specifically include: first, the rectangular frame label of the tank top is obtained from the original tank top detection photo by means of manual frame selection; after the rectangular frame selection, the rust label and the tank top mask label can be further generated by means of automatic and semi-automatic labeling.

[0107] In practical applications, the specific steps of generating the rust label by means of the automatic labeling generation method can include:

[0108] The rust seed region features are selected by means of the clustering method;

[0109] The preliminary rust label is obtained by means of the seed region growth method;

[0110] The central region features are updated by means of clustering again;

[0111] The rust label is updated by means of the seed region growth method again;

[0112] The process is ended when the clustering center no longer changes or reaches the iteration termination signal.

[0113] In practical applications, the specific steps of generating the tank top mask label by means of the semi-automatic labeling generation method can be as follows:

[0114] The edge detection selects the tank top candidate region;

[0115] The largest polygon object is selected as the coarse label;

[0116] The final tank top mask label is obtained after manual fine-tuning.

[0117] It should be noted that the tank top detection photo or the original tank top detection photo in the embodiment of the present application can be preprocessed, such as extracting the illumination field, to eliminate the influence and interference of different illumination environments on the brightness of the photo.

[0118] S33, calculate the tank top rust area ratio of the storage tank by counting the rust pixels of the tank top image;

[0119] In the embodiment of the present application, after the rust on the tank top of the storage tank is identified, the tank top rust area ratio of the storage tank can be calculated by counting the rust pixels in the tank top image.

[0120] S34, generate the tank top detection result of the tank top of the storage tank according to the tank top rust area ratio.

[0121] In actual application, the tank top detection result corresponding to the detected tank can be generated according to the tank top rust area ratio according to the corresponding relationship between the rust area ratio and the rust degree.

[0122] Further, in the embodiment of the present application, the steps further include:

[0123] S41, the detection data further includes tank top scanning data; a tank top SAR image consistent with a tank top detection photo coordinate system is generated according to the tank top scanning data corresponding to the tank top scanning data; the unmanned aerial vehicle is further provided with a synthetic aperture radar for obtaining tank top scanning data synchronized with the image acquisition device; the tank top scanning data is synchronized and one-to-one corresponding to the tank top detection photo obtained by the image acquisition device;

[0124] If the unmanned aerial vehicle carries a synthetic aperture radar alone to scan and obtain tank top scanning data of the tank area, the position accuracy deviation caused by the spatial inertial navigation combination solution of the moving synthetic aperture radar will make the SAR image obtained thereby unable to be directly used; therefore, in the embodiment of the present application, the unmanned aerial vehicle carries an image acquisition device and a synthetic aperture radar simultaneously, and synchronously obtains tank top detection photos and tank top scanning data; thus, referring to the coordinates of the tank top detection photo corresponding to the tank top scanning data to generate a tank top SAR image can avoid the problem of position accuracy deviation of the tank top SAR image generated by the unmanned aerial vehicle carrying a synthetic aperture radar alone.

[0125] S42, according to the rust pixels of the tank top image, a corresponding tank top reflected rust mark area is determined in the tank top SAR image;

[0126] Since the coordinates of the tank top SAR image and the coordinates of the inspection photo are consistent, in the embodiment of the present application, the rust mark area in the rust mark recognition result of the tank top image can be determined as a tank top reflected rust mark area from the tank top SAR image according to the rust mark recognition model of the tank top image;

[0127] S43, according to the gray value of the tank top reflected rust mark area, the rust depth of the tank top reflected rust mark area is determined.

[0128] In actual application, when the rust depth of the reflected rust mark area is determined, the corresponding relationship between the gray value of the tank top SAR image and the rust depth can be preset; then the corresponding rust depth is determined according to the gray value of the reflected rust mark area in the tank top SAR image and the corresponding relationship.

[0129] In conclusion, the rust on the tank roof can be accurately identified by the application, so that the rust area ratio of the tank roof can be accurately calculated according to the identification result; and then the accurate result of the rust determination of the tank can be obtained according to the rust area ratio; since the application eliminates the subjective judgment link, the determination result is more accurate and stable.

[0130] Further, in the application, the tank roof scanning data collected by the synthetic aperture radar synchronized with the image collection device can also be included, and the tank roof SAR image consistent with the corresponding inspection photo coordinate system is generated according to the tank roof scanning data; since the scattering characteristics of the tank roof SAR image can reflect the roughness of the surface of the photographed object, the rust area can be determined in the tank roof SAR image; then, the corresponding rust depth is determined according to the different scattering characteristics in the rust area; since the application can not only determine the rust area (or rust area ratio) of the tank roof, but also identify the different rust depths in the rust area, the determination result of the rust degree and rust condition can be more accurate and comprehensive.

[0131] Embodiment four

[0132] In another aspect of the embodiment of the application, a rust detection device for a storage tank is also provided, Figure 3 The structure of the rust detection device for a storage tank provided by the embodiment of the application is shown in the schematic diagram, and the rust detection device for a storage tank is a virtual device Figure 1 Or Figure 2 The system corresponding to the rust detection method for a storage tank in the corresponding embodiment, that is, the virtual device is realized Figure 1 Or Figure 2 The rust detection method for a storage tank in the corresponding embodiment constitutes each virtual module of the rust detection device for a storage tank, which can be executed by an electronic device, such as a network device, a terminal device, or a server. The rust detection device for a storage tank in the embodiment of the application can realize the rust detection for a storage tank required by industrial control. Specifically, the rust detection device for a storage tank in the embodiment of the application includes:

[0133] The tank wall picture generation unit 01 is used to establish the tank wall three-dimensional image of the detected storage tank according to the detection data including the three-dimensional point cloud data of the tank wall of the storage tank, and generate the first side development diagram of the tank wall surface of the detected storage tank according to the tank wall three-dimensional image by picture stitching; the three-dimensional point cloud data is obtained by a three-dimensional laser scanner;

[0134] The tank wall rust identification unit 02 is used to identify rust on the first side view unfolded image by using a tank wall rust identification model based on an artificial neural network, and to determine the rust pixels from the first side view unfolded image; the tank wall rust identification model is constructed by using previously generated first side view unfolded images of the tank wall as modeling data and is used to identify rust on the tank wall.

[0135] Tank wall corrosion ratio calculation unit 03 is used to calculate the tank wall corrosion area ratio by statistically analyzing the corrosion pixels in the first side view unfolded image; the tank wall corrosion area ratio is the ratio of the corrosion area to the area of ​​the first side view unfolded image.

[0136] Tank wall inspection result generation unit 04 is used to generate tank wall inspection results based on the ratio of rust area to tank wall.

[0137] Furthermore, in embodiments of the present invention, it may also include:

[0138] The tank wall SAR image generation unit 11 includes tank wall scanning data, which is used to create a three-dimensional SAR image of the tank wall of the detected storage tank based on the tank wall scanning data, and to obtain a second side view of the curved surface of the tank wall of the detected storage tank by image stitching based on the three-dimensional SAR image of the tank wall; the tank wall scanning data is obtained by scanning the tank wall of the storage tank with a synthetic aperture radar located at the same position as the three-dimensional laser scanner.

[0139] The tank wall SAR image adjustment unit 12 converts the area size and coordinate system of the second side unfolded image to be consistent with the first side unfolded image. Figure One To;

[0140] The tank wall rust area determination unit 13 is used to determine the corresponding tank wall reflective rust area in the second side unfolded image based on the rust pixels in the first side unfolded image.

[0141] The tank wall corrosion depth determination unit 14 is used to determine the corrosion depth of the tank wall reflective rust area based on the gray value of the reflective rust area of ​​the tank wall.

[0142] Because the working principle and beneficial effects of the rust detection device for storage tanks in the embodiments of the present invention have already been demonstrated... Figure 1 The corresponding methods for detecting rust in storage tanks are also described and explained, so they can be referenced together and will not be repeated here.

[0143] Example 5

[0144] Corresponding to the method embodiments, the application also provides a rust detection device for a storage tank, such as a terminal, a server, etc. The server can be a physical server, a server cluster or a distributed system formed by multiple physical servers, a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms. The terminal can be a smartphone, a tablet computer, a notebook computer, a desktop computer, etc., but is not limited thereto.

[0145] The example diagram of the hardware structure block diagram of the rust detection device for the storage tank provided by the embodiments of the application is shown in Figure 4 as shown, which can include:

[0146] a processor 1, a communication interface 2, a memory 3, and a communication bus 4;

[0147] The processor 1, the communication interface 2, and the memory 3 can communicate with each other through the communication bus 4.

[0148] Optionally, the communication interface 2 can be an interface of a communication module, such as an interface of a GSM module.

[0149] The processor 1 can be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the application.

[0150] The memory 3 can include a high-speed RAM memory, and can also include a non-volatile memory, such as at least one disk memory.

[0151] The processor 1 is specifically configured to execute a computer program stored in the memory 3 to perform the following steps:

[0152] S11, establishing a tank wall three-dimensional image of a detected storage tank according to detection data including three-dimensional point cloud data of a tank wall of the storage tank, and generating a first side development drawing of a tank wall surface of the detected storage tank according to the tank wall three-dimensional image through picture stitching; the three-dimensional point cloud data is obtained through a three-dimensional laser scanner;

[0153] S12, rust mark recognition is performed on the first side development drawing through a tank wall rust mark recognition model based on an artificial neural network, and rusted pixels are determined from the first side development drawing; the tank wall rust mark recognition model is constructed through data training by taking the first side development drawing of the tank wall of the storage tank generated in the past as modeling data, and is used for recognizing rust marks on the tank wall of the storage tank;

[0154] S13, calculate a tank wall corrosion area ratio of the tank wall of the storage tank by counting the corrosion pixels of the first side unfolding map; the tank wall corrosion area ratio is a ratio of a corrosion area to an area of the first side unfolding map;

[0155] S14, generate a tank wall detection result of the tank wall of the storage tank according to the tank wall corrosion area ratio.

[0156] Preferably, in the embodiments of the present application, the method further comprises the steps of:

[0157] S21, the detection data further comprises tank wall scanning data, a three-dimensional SAR image of the tank wall of the detected storage tank is established according to the tank wall scanning data, and a second side unfolding map of the tank wall surface of the detected storage tank is obtained by picture stitching according to the three-dimensional SAR image of the tank wall; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner scanning the tank wall;

[0158] S22, convert the area size and coordinate system of the second side unfolding map into the first side unfolding map; Figure One

[0159] S23, determine a corresponding tank wall reflection rust area in the second side unfolding map according to the corrosion pixels in the first side unfolding map;

[0160] S24, determine the corrosion depth of the tank wall reflection rust area according to the gray value of the tank wall reflection rust area.

[0161] Preferably, in the embodiments of the present application, the method further comprises the steps of:

[0162] S31, the detection data further comprises a tank top detection photo; a tank top image of the detected storage tank is generated according to the tank top detection photo; the tank top detection photo is obtained by an image acquisition device arranged on a drone; the tank top detection photo is an orthographic image;

[0163] S32, rust mark recognition is performed on the tank top image by a tank top rust mark recognition model based on an artificial neural network, and corrosion pixels are determined from the tank top image; the tank top rust mark recognition model is constructed by taking the tank top images of the tank tops of the storage tanks generated in the past as modeling data and through data training, and is used for recognizing rust marks on the tank top of the storage tank;

[0164] S33, calculate a tank top corrosion area ratio of the tank top of the storage tank by counting the corrosion pixels of the tank top image;

[0165] S34, generate a tank top detection result of the tank top of the storage tank according to the tank top corrosion area ratio.

[0166] ​Preferably, in the embodiments of the present application, the method can further comprise the steps of:

[0167] S41, the detection data further comprises tank top scanning data; a tank top SAR image corresponding to the tank top scanning data is generated according to the tank top scanning data; the unmanned aerial vehicle is further provided with a synthetic aperture radar for acquiring tank top scanning data synchronized with the image acquisition device; the tank top scanning data is synchronized with and one-to-one corresponds to the tank top detection photo acquired by the image acquisition device;

[0168] S42, according to the rust pixels of the tank top image, a corresponding tank top reflected rust mark area is determined in the tank top SAR image;

[0169] S43, according to the gray value of the tank top reflected rust mark area, the rust depth of the tank top reflected rust mark area is determined.

[0170] The computer program product for the corrosion detection device for the storage tank in the embodiments of the present application comprises program instructions which, when executed by a computer, can enable the computer to execute the corrosion detection method for the storage tank described in the above aspects and achieve the same technical effects.

[0171] Embodiment six

[0172] In the embodiments of the present application, a storage medium is also provided, which can store a program suitable for execution by a processor, and the program is used for:

[0173] S11, a tank wall three-dimensional image of a detected storage tank is established according to detection data comprising three-dimensional point cloud data of a tank wall of the storage tank, and a first side development drawing of a tank wall surface of the detected storage tank is generated according to the tank wall three-dimensional image through picture stitching; the three-dimensional point cloud data is obtained by a three-dimensional laser scanner;

[0174] S12, rust mark recognition is performed on the first side development drawing through a tank wall rust mark recognition model based on an artificial neural network, and rust pixels are determined from the first side development drawing; the tank wall rust mark recognition model is constructed through data training by taking the first side development drawing of the tank wall of the storage tank generated in the past as modeling data, and is used for recognizing rust marks on the tank wall of the storage tank;

[0175] S13, a tank wall rust area ratio of the tank wall of the storage tank is calculated by counting the rust pixels of the first side development drawing; the tank wall rust area ratio is a ratio of a rust area to an area of the first side development drawing;

[0176] S14, a tank wall detection result of the tank wall of the storage tank is generated according to the tank wall rust area ratio.

[0177] Preferably, in the embodiments of the present application, the method can further comprise the steps of:

[0178] S21, the detection data further comprises tank wall scanning data, a three-dimensional SAR image of the tank wall of the detected storage tank is established according to the tank wall scanning data, and a second side development drawing of the tank wall surface of the detected storage tank is obtained according to the three-dimensional SAR image of the tank wall; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner and scanning the tank wall of the storage tank;

[0179] S22, the area size and coordinate system of the second side development drawing are converted into the first side development drawing Figure One ;

[0180] S23, according to the rust pixels in the first side development drawing, the corresponding tank wall reflection rust area in the second side development drawing is determined;

[0181] S24, according to the gray value of the tank wall reflection rust area, the rust depth of the tank wall reflection rust area is determined.

[0182] Preferably, in the embodiment of the present application, the steps can further include:

[0183] S31, the detection data further comprises tank top detection photos; the tank top image of the detected storage tank is generated according to the tank top detection photos; the tank top detection photos are obtained by the image acquisition device arranged on the unmanned aerial vehicle; the tank top detection photos are orthographic images;

[0184] S32, rust mark recognition is performed on the tank top image by a tank top rust mark recognition model based on artificial neural network, and rust pixels are determined from the tank top image; the tank top rust mark recognition model is constructed by taking the tank top images of the tank top of the storage tank generated in the past as modeling data and through data training, and is used for identifying rust marks on the tank top of the storage tank;

[0185] S33, the tank top rust area ratio of the tank top of the storage tank is calculated by counting the rust pixels of the tank top image;

[0186] S34, the tank top detection result of the tank top of the storage tank is generated according to the tank top rust area ratio.

[0187] Preferably, in the embodiment of the present application, the steps can further include:

[0188] S41, the detection data further comprises tank top scanning data; a tank top SAR image consistent with the tank top detection photo coordinate system corresponding to the tank top scanning data is generated according to the tank top scanning data; the unmanned aerial vehicle is further provided with a synthetic aperture radar for obtaining tank top scanning data synchronous with the image acquisition device; the tank top scanning data is synchronous and one-to-one corresponding with the tank top detection photos obtained by the image acquisition device;

[0189] S42, determining a corresponding tank top reflective rust area in the tank top SAR image according to the rust pixels of the tank top image;

[0190] S43, determining a rust depth of the tank top reflective rust area according to the gray value of the tank top reflective rust area.

[0191] Optionally, the refinement function and the extension function of the program can refer to the above description.

[0192] The product described above can execute the method provided by the embodiments of the application, and has the corresponding function modules and beneficial effects of the execution method. Technical details not described in detail in the embodiments can refer to the method provided by the embodiments of the application.

[0193] Those skilled in the art can appreciate that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0194] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be realized by other ways. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0195] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0196] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0197] It should be understood that the features in the embodiments of the present application, each embodiment, and features can be combined with each other, and can achieve the solution to the above technical problems.

[0198] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the parts that contribute to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0199] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting corrosion of a storage tank, characterized by, The method comprises the steps of: S11, establishing a tank wall three-dimensional image of a detected tank according to detection data comprising three-dimensional point cloud data of the tank wall, and generating a first side development drawing of a tank wall surface of the detected tank according to the tank wall three-dimensional image through picture stitching; The three-dimensional point cloud data is obtained by a three-dimensional laser scanner; S12, rust mark recognition is performed on the first side development drawing through a tank wall rust mark recognition model based on an artificial neural network, and rust pixels are determined from the first side development drawing; the tank wall rust mark recognition model is constructed through data training by taking previously generated first side development drawings of tank walls as modeling data, and is used for recognizing rust marks on tank walls; S13, a tank wall rust area ratio of the tank wall is calculated by counting rust pixels of the first side development drawing; the tank wall rust area ratio is a ratio of a rust area to an area of the first side development drawing; S14, a tank wall detection result of the tank wall is generated according to the tank wall rust area ratio; Further comprising the steps of: S21, the detection data further comprises tank wall scanning data, a three-dimensional SAR image of the tank wall of the detected tank is established according to the tank wall scanning data, and a second side development drawing of the tank wall surface of the detected tank is obtained according to the three-dimensional SAR image of the tank wall through picture stitching; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner; S22, the area size and coordinate system of the second side development drawing are converted to be consistent with the first side development drawing; S23, according to the rust pixels in the first side development drawing, a corresponding tank wall reflection rust mark area is determined in the second side development drawing; S24, according to the gray value of the tank wall reflection rust mark area, the rust depth of the tank wall reflection rust mark area is determined, comprising: a preset corresponding relationship between the gray value of the SAR image and the rust depth; determining the rust depth of the tank wall reflection rust mark area according to the gray value of the tank wall reflection rust mark area and the corresponding relationship.

2. The corrosion detection method for a storage tank according to claim 1, characterized by, Further comprising: dividing the tank wall reflection rust mark area according to the rust depth of the tank wall reflection rust mark area.

3. The corrosion detection method for a storage tank according to claim 2, characterized by, The dividing of the tank wall reflection rust mark area according to the rust depth of the tank wall reflection rust mark area comprises: a preset corresponding relationship between different rust depth ranges and rust degrees; determining the area or proportion of each rust degree in the tank wall reflection rust mark area according to the corresponding relationship.

4. The corrosion detection method for a storage tank according to claim 3, characterized by, Further comprising: S31, the detection data further comprises a tank top detection photo; a tank top image of the detected tank is generated according to the tank top detection photo; the tank top detection photo is obtained by an image acquisition device arranged on a drone; the tank top detection photo is an orthographic image; S32, rust mark recognition is performed on the tank top image through a tank top rust mark recognition model based on an artificial neural network, and rust pixels are determined from the tank top image; the tank top rust mark recognition model is constructed through data training by taking previously generated tank top images of tank tops as modeling data, and is used for recognizing rust marks on tank tops; S33, calculate a tank top rust area ratio of the tank top by counting rust pixels of the tank top image; S34, generate a tank top detection result of the tank top according to the tank top rust area ratio.

5. The corrosion detection method for a storage tank according to claim 4, characterized by, Further comprising: S41, the detection data further comprises tank top scanning data; a tank top SAR image corresponding to the tank top scanning data is generated according to the tank top scanning data in a tank top detection photo coordinate system; the unmanned aerial vehicle is further provided with a synthetic aperture radar for acquiring tank top scanning data synchronized with the image acquisition device; the tank top scanning data is synchronized and one-to-one corresponding with the tank top detection photo acquired by the image acquisition device; S42, according to the rust pixels of the tank top image, determine the corresponding tank top reflected rust mark area in the tank top SAR image; S43, according to the gray value of the tank top reflected rust mark area, determine the rust depth of the tank top reflected rust mark area.

6. The corrosion detection method for a storage tank according to claim 5, characterized by, The tank top rust mark recognition model comprises: A rectangular recognition sub-model for generating a rectangular frame label of the detected tank top in the tank top detection photo; the rectangular recognition sub-model is trained by combining the original tank top detection photo with the rectangular frame label to generate; the rectangular frame label of the tank top pattern is obtained by framing the original tank top detection photo as historical data, and a tank top mask label and a rust mark label are generated; A mask recognition sub-model for generating a tank top mask label of the detected tank in the tank top rectangular image; the mask recognition sub-model is trained by combining the tank top rectangular image with the tank top mask label to generate; the tank top rectangular image is obtained according to the rectangular frame label; A preliminary rust mark recognition sub-model for obtaining a preliminary rust mark recognition result according to the tank top rectangular image; the preliminary rust mark recognition sub-model is trained by combining the tank top rectangular image with the rust mark label to generate; A tank top recognition result module for determining rust pixels from the tank top image according to the tank top mask label generated by the mask recognition sub-model and the preliminary rust mark recognition result.

7. A corrosion detection apparatus for a storage tank, characterized by, Comprising: A tank wall picture generation unit for establishing a tank wall three-dimensional image of the detected tank according to detection data comprising three-dimensional point cloud data of the tank wall of the tank, and generating a first side development drawing of a tank wall surface of the detected tank according to the tank wall three-dimensional image by picture stitching; The three-dimensional point cloud data is obtained by a three-dimensional laser scanner; A tank wall rust mark recognition unit for determining rust pixels from the first side development drawing by a tank wall rust mark recognition model based on artificial neural network; the tank wall rust mark recognition model is constructed by data training by taking the first side development drawing of the tank wall of the tank generated in the past as modeling data, and is used for recognizing rust marks on the tank wall; A tank wall rust area ratio calculation unit for calculating a tank wall rust area ratio of the tank wall by counting rust pixels of the first side development drawing; the tank wall rust area ratio is the ratio of the rust area to the area of the first side development drawing; The tank wall detection result generation unit is configured to generate a tank wall detection result of the tank wall of the storage tank according to the tank wall rust area ratio; The tank wall SAR image generation unit is configured to generate a three-dimensional SAR image of the tank wall of the detected storage tank according to the tank wall scanning data, and obtain a second side development drawing of the tank wall surface of the detected storage tank by picture stitching according to the three-dimensional SAR image of the tank wall; the tank wall scanning data is obtained by a synthetic aperture radar arranged at the same position as the three-dimensional laser scanner and scanning the tank wall of the storage tank; The tank wall SAR image adjustment unit is configured to convert the area size and coordinate system of the second side development drawing to be consistent with the first side development drawing; The tank wall rust area determination unit is configured to determine a corresponding tank wall reflection rust area in the second side development drawing according to the rust pixels in the first side development drawing; The tank wall rust depth determination unit is configured to determine the rust depth of the tank wall reflection rust area according to the gray value of the tank wall reflection rust area, including: a preset corresponding relationship between the gray value of the SAR image and the rust depth; and determining the rust depth of the tank wall reflection rust area according to the gray value of the tank wall reflection rust area and the corresponding relationship.

8. A rust detection device for a storage tank, comprising: a memory configured to store a computer program; a processor configured to call and execute the computer program to implement each step of the rust detection method for a storage tank according to any one of claims 1-6.

9. A storage medium having a computer program stored thereon, the computer program being executed by a processor to implement each step of the rust detection method for a storage tank according to any one of claims 1-6.

Images