Method, device and equipment for underwater target recognition of multiple obstacle detectors based on manual calibration
By introducing artificial calibration mechanisms and multi-failer collaboration mechanisms into the underwater target recognition system, the problems of target recognition accuracy and robustness in complex underwater environments are solved, and higher recognition accuracy and system adaptability are achieved.
Patent Information
- Application Number
- CN202411625218.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-14
- Publication Date
- 2025-04-04
AI Technical Summary
In complex underwater environments, it is difficult for the prior art to achieve high accuracy and robust underwater target recognition, especially in cases of multi-objective dense, complex background and illumination variation.
The multi-failer underwater target recognition method based on manual calibration is adopted. By receiving the real-time video stream of the detector, obtaining external box selection input and target type and confidence information, the target latitude and longitude of the target is calculated, and the recognition accuracy is improved through secondary cross-confirmation and model update.
It significantly improves the accuracy and robustness of target recognition, enhances the flexibility and adaptability of the system, and can achieve more accurate and reliable target recognition in complex underwater environments.
Smart Images
Figure CN119762959A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of underwater target recognition, and in particular to a method, device, equipment and medium for underwater target recognition of multiple obstacle detectors based on manual calibration. Background Art
[0002] Underwater target recognition technology plays an increasingly important role in many fields such as ocean exploration, military applications, and environmental monitoring. The underwater target recognition method based on the obstacle detector automatically recognizes the target in the picture based on the image or video stream collected in the monitoring area through the built-in recognition algorithm, and transmits the captured image or video stream and the target recognition result to the control host for the operator to view and further process.
[0003] The target recognition systems in related technologies mostly use pre-trained deep learning models for target detection and recognition, such as convolutional neural networks (CNNs). These models need to be trained with a large amount of labeled data to obtain good recognition performance. Although they perform well in some specific scenarios, in some complex and changeable underwater environments, the complexity of the underwater environment far exceeds the scope of the training data set, resulting in a significant decrease in recognition accuracy when the model faces multiple dense targets, complex backgrounds, and changing lighting. At the same time, this limitation not only increases the risk of misidentification and missed identification, but also directly affects the reliability and practicality of the system. Summary of the invention
[0004] The present invention provides a multi-obstacle detector underwater target recognition method, device, equipment and medium based on manual calibration, which solves the problem of limited accuracy and robustness of underwater target recognition in complex underwater environments.
[0005] In order to achieve the above objectives, this application adopts the following technical solutions:
[0006] In a first aspect, a method for underwater target recognition of multiple obstacle detectors based on manual calibration is provided, comprising:
[0007] Receive and decode the real-time video streams of multiple obstacle detectors and display them in the form of images;
[0008] For an unrecognized target, an external frame selection input is obtained, and the unrecognized target is calibrated by the frame selection input; at the same time, an external target type T1 and confidence C1 input are obtained;
[0009] Calculate the longitude and latitude of the calibration target based on the center point of the frame selection input;
[0010] Sending the latitude and longitude of the calibration target, the target type T1 and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system;
[0011] The updated recognition results are continuously sent to the model training modules of the plurality of obstacle detectors to update the recognition model.
[0012] In a first possible implementation manner of the first aspect, calculating the longitude and latitude of the calibration target based on the center point of the box selection input includes:
[0013] Based on the known obstacle detector coordinate system, calculating the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system;
[0014] Based on the known north angle of the obstacle detector, converting the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system into the distance relative to the obstacle detector in the UTM coordinate system;
[0015] Based on the UTM coordinate system projection method, the known longitude and latitude of the obstacle detector are converted into coordinates in the UTM coordinate system;
[0016] According to the coordinates of the obstacle detector in the UTM coordinate system, the distance of the calibration target relative to the obstacle detector is superimposed in the X direction and the Y direction to obtain the coordinates of the calibration target in the UTM coordinate system;
[0017] According to the coordinates of the calibration target in the UTM coordinate system, the coordinate values of the calibration target in the UTM coordinate system are converted into longitude and latitude using the UTM coordinate system projection method.
[0018] Based on the first possible implementation manner of the first aspect, in a second possible implementation manner of the first aspect, sending the latitude and longitude of the calibration target, the target type T1, and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtaining a secondary recognition result, and updating the recognition result of the system includes:
[0019] The obstacle detector is configured to perform secondary recognition on the calibration target according to the acquired latitude and longitude of the calibration target, and to provide the target type T2 and confidence C2 according to the built-in recognition algorithm;
[0020] According to the comparison, when the target type T1 and the target type T2 are consistent, the target type is determined to be T1, and the final confidence is (C1+C2) / 2;
[0021] When the target type T1 and the target type T2 are inconsistent, the two detection results are saved locally, and wait for external input to be obtained to make a final determination and update the recognition result.
[0022] Based on the second possible implementation of the first aspect, in a third possible implementation of the first aspect, if the obstacle detector cannot identify the calibration target during secondary identification, the recognition result of the system is updated according to externally input manual calibration data.
[0023] In a second aspect, a multi-obstacle detector underwater target recognition device based on manual calibration is provided, comprising:
[0024] A data acquisition module is used to receive and decode the real-time video streams of multiple obstacle detectors and display them in the form of images;
[0025] A calibration module is used to obtain an external frame selection input for an unrecognized target, and calibrate the unrecognized target through the frame selection input; at the same time, obtain an external target type T1 and confidence C1 input;
[0026] A latitude and longitude calculation module, used to calculate the longitude and latitude of the calibration target based on the center point of the box selection input;
[0027] A secondary cross-confirmation module is used to send the latitude and longitude of the calibration target, the target type T1 and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system;
[0028] The recognition model updating module is used to continuously send the updated recognition results to the model training modules of the multiple obstacle detectors to update the recognition model.
[0029] In a first possible implementation manner of the second aspect, the latitude and longitude calculation module is specifically used to:
[0030] Based on the known obstacle detector coordinate system, calculating the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system;
[0031] Based on the known north angle of the obstacle detector, converting the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system into the distance relative to the obstacle detector in the UTM coordinate system;
[0032] Based on the UTM coordinate system projection method, the known longitude and latitude of the obstacle detector are converted into coordinates in the UTM coordinate system;
[0033] According to the coordinates of the obstacle detector in the UTM coordinate system, the distance of the calibration target relative to the obstacle detector is superimposed in the X direction and the Y direction to obtain the coordinates of the calibration target in the UTM coordinate system;
[0034] According to the coordinates of the calibration target in the UTM coordinate system, the coordinate values of the calibration target in the UTM coordinate system are converted into longitude and latitude using the UTM coordinate system projection method.
[0035] In a second possible implementation manner of the second aspect, the secondary cross-confirmation module is specifically configured to:
[0036] The obstacle detector is configured to perform secondary recognition on the calibration target according to the acquired latitude and longitude of the calibration target, and to provide the target type T2 and confidence C2 according to the built-in recognition algorithm;
[0037] According to the comparison, when the target type T1 and the target type T2 are consistent, the target type is determined to be T1, and the final confidence is (C1+C2) / 2;
[0038] When the target type T1 and the target type T2 are inconsistent, the two detection results are saved locally, and wait for external input to be obtained to make a final determination and update the recognition result.
[0039] Based on the second possible implementation manner of the second aspect, in a third possible implementation manner of the second aspect, the secondary cross-confirmation module is further used to:
[0040] If the obstacle detector cannot identify the calibration target during the secondary identification, the recognition result of the system is updated according to the manual calibration data input externally.
[0041] According to a third aspect, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the multi-obstacle detector underwater target identification method based on manual calibration as described in the first aspect are implemented.
[0042] In a fourth aspect, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the multi-obstacle detector underwater target identification method based on manual calibration as described in the first aspect are implemented.
[0043] This application significantly improves the accuracy, flexibility and adaptability of the target recognition and detection system by enhancing the manual calibration function, adopting a multi-obstacle detector collaboration mechanism, and realizing dynamic adjustment and adaptive optimization of the model. It can meet the needs of wide-area monitoring and multi-target accurate recognition and positioning in complex scenarios, and overcomes the shortcomings of existing technologies in target recognition accuracy, coverage range and system sustainable optimization capabilities. It can achieve more accurate and reliable target recognition in complex and changeable underwater environments by combining the operator's intuitive judgment and the machine's automatic processing. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 A schematic flow chart of a multi-obstacle detector underwater target recognition method based on manual calibration provided in an embodiment of the present application;
[0045] Figure 2 A schematic diagram of a box selection input provided in an embodiment of the present application;
[0046] Figure 3 A schematic diagram of the conversion between a calibration target obstacle detector coordinate system and a UTM coordinate system provided in an embodiment of the present application;
[0047] Figure 4 A structural block diagram of a multi-obstacle detector underwater target recognition device based on manual calibration provided in an embodiment of the present application;
[0048] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0049] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose, the technical solutions in the embodiments of the present application are clearly described. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present application.
[0050] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.
[0051] The description of the method flow in the specification of the present application and the steps of the flowchart in the drawings of the present specification do not have to be strictly executed according to the step numbers, and the method steps can be executed in a different order. Moreover, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps.
[0052] The following is a detailed description of the multi-obstacle detector underwater target recognition method, device, equipment and medium based on manual calibration provided in the embodiments of the present application in combination with the accompanying drawings and preferred embodiments.
[0053] First, the application scenario of the multi-obstacle detector underwater target recognition method based on manual calibration in an embodiment of the present application is described in detail.
[0054] Due to the complex underwater environment, insufficient light, low visibility, and high noise interference, especially in wide-area monitoring and multi-target environments, how to improve recognition accuracy and positioning precision has become a major challenge facing existing technologies.
[0055] The target recognition systems in related technologies mostly rely on pre-trained models for target recognition, but in complex scenarios (such as dense multi-targets, complex underwater backgrounds, changing lighting, etc.), the accuracy of recognition is often difficult to guarantee. This limitation makes the system prone to misidentification or missed recognition in certain scenarios, which in turn affects the reliability of the entire system.
[0056] Moreover, the target recognition system is highly dependent on automated processing procedures and lacks the function of real-time human intervention and correction. In particular, when target omissions and false alarms occur, there is a lack of effective real-time human intervention mechanisms, and the automated processing procedures cannot respond to operator feedback in a timely manner. Once an identification error occurs, it is difficult for operators to intervene quickly and make corrections. This not only limits the flexibility and adaptability of the system, but may also have an adverse impact on subsequent decision-making and operations.
[0057] Based on this, this application provides a multi-obstacle detector underwater target recognition method based on manual calibration, which aims to improve the accuracy and robustness of underwater target recognition, while enhancing the flexibility and adaptability of the system. By introducing a manual calibration mechanism, the recognition parameters can be dynamically adjusted in a complex environment, effectively reducing the occurrence of misidentification and missed recognition, thereby improving the overall performance of the system.
[0058] See also Figure 1 , the embodiment of the present application provides a multi-obstacle detector underwater target recognition method based on manual calibration, such as Figure 1 As shown, the method of the embodiment of the present application includes the following steps:
[0059] Step S1, receiving and decoding real-time video streams of multiple obstacle detectors, and displaying them in the form of images.
[0060] In specific implementation, the control host receives real-time video streams from multiple obstacle detectors through the network or a dedicated communication link. After being decoded, these video streams are displayed on the display screen of the control host in the form of images. The video stream of each obstacle detector contains different viewing angles and position information.
[0061] Step S2, for the unrecognized target, obtain the external box selection input, and calibrate the unrecognized target through the box selection input; at the same time, obtain the external target type T1 and confidence C1 input.
[0062] During the specific implementation process, when an unidentified target is found, the operator can use the mouse on the control host to calibrate the target on the image, and manually specify the target type T1 (for example, fish, divers, ship wreckage, etc.) and confidence C1 (indicating the operator's confidence in the target type, which can be a value between 0 and 1).
[0063] Step S3, based on the center point of the frame selection input, calculate the longitude and latitude of the calibration target.
[0064] This step calculates the longitude and latitude of the calibrated target based on the input center point, UTM coordinate system projection method and the longitude and latitude of the obstacle detector, that is, calculates the global coordinate system of the manually calibrated target position for accurate positioning and navigation. Specifically, it includes:
[0065] Step S301, based on the known obstacle detector coordinate system, calculate the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system.
[0066] Step S302: based on the known north angle of the obstacle detector, convert the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system into the distance relative to the obstacle detector in the UTM coordinate system.
[0067] For example, see Figure 2-3 , the calibration frame is a rectangle, and the distance (Δx, Δy) of the target relative to the obstacle detector in the obstacle detector's own coordinate system is calculated using the center point of the rectangle; based on the obstacle detector's north angle θ, the distance (Δx', Δy') of the calibration target relative to the obstacle detector in the UTM (Universal Transverse Mercator) rectangular coordinate system is calculated. The conversion relationship expression is:
[0068]
[0069] Step S303, based on the UTM coordinate system projection method, convert the known longitude and latitude of the obstacle detector into coordinates in the UTM coordinate system.
[0070] Use the UTM coordinate system projection method to convert the longitude and latitude of the obstacle detector into coordinates (x0, y0) in the UTM rectangular coordinate system.
[0071] Step S304: According to the coordinates of the obstacle detector in the UTM coordinate system, the distance of the calibration target relative to the obstacle detector is superimposed in the X direction and the Y direction respectively to obtain the coordinates of the calibration target in the UTM coordinate system.
[0072] According to (x0, y0), the distance (Δx', Δy') from the target to the obstacle detector is superimposed in the X direction and the Y direction respectively to obtain the position (x1, y1) of the target in the UTM rectangular coordinate system.
[0073] Step S305 , according to the coordinates of the calibration target in the UTM coordinate system, the coordinate values of the calibration target in the UTM coordinate system are converted into longitude and latitude using the UTM coordinate system projection method.
[0074] Step S4, sending the latitude and longitude of the calibration target, target type T1 and confidence C1 to one of the obstacle detectors that is not in the detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system.
[0075] One of the obstacle detectors obtains the latitude and longitude, target type T1 and confidence C1 of the calibration target, and during the non-detection mission period, performs a secondary cross-confirmation on the calibration target and updates the recognition result to the control host.
[0076] Step S4 specifically includes the following steps:
[0077] Step S401, configure the obstacle detector to perform secondary recognition on the calibration target according to the acquired latitude and longitude of the calibration target, and give the target type T2 and confidence C2 according to the built-in recognition algorithm.
[0078] Step S402: According to the comparison, when the target type T1 and the target type T2 are consistent, the target type is determined to be T1, and the final confidence is (C1+C2) / 2.
[0079] Step S403: when the target type T1 and the target type T2 are inconsistent, the two detection results are saved locally, and wait for external input to be obtained to make a final determination and update the recognition result.
[0080] Exemplarily, the system includes a main obstacle detector and a secondary obstacle detector. In specific implementation, after the operator calibrates the target manually in the main obstacle detector, the operator controls the host to send the manually calibrated target latitude and longitude, type T1 and confidence T1 information to the secondary obstacle detector through a communication link. The secondary obstacle detector performs a secondary cross-confirmation of the manually calibrated target during idle time when not performing tasks to ensure the accuracy of identification and positioning.
[0081] If the target type T2 identified by the secondary obstacle detector is the same as the target type T1 calibrated by the operator, the system will confirm that the target type is T1, and the final confidence level is the average of C1 and C2. This can comprehensively consider the confidence levels of the operator and the secondary obstacle detector and improve the reliability of identification.
[0082] If the target type T2 identified by the secondary obstacle detector is inconsistent with the target type T1 calibrated by the operator, the system will save the two detection results locally and notify the operator for manual confirmation. The operator can manually select the correct target type according to the actual situation and update the system's recognition results.
[0083] In some possible implementations, if the obstacle detector cannot identify the calibration target during secondary identification, the system's identification result is updated based on externally input manual calibration data.
[0084] Step S5, continuously sending the updated recognition results to the model training modules of multiple obstacle detectors to update the recognition model.
[0085] The control host sends the final confirmed target image, target type, confidence and other information to the model training modules of the main and auxiliary obstacle detectors. These modules use the new recognition results as training data to update and optimize the existing recognition models. Through continuous learning and iteration, the system can gradually improve the accuracy and robustness of target recognition.
[0086] Based on the above technical solution, the present application provides a manual calibration function in which an operator can intervene, allowing the operator to perform real-time target calibration on the image of the obstacle detector, convert the pixel coordinates of the target in the image into the relative distance from the obstacle detector, and combine the longitude and latitude position information of the obstacle detector. The UTM coordinate system projection method is used to accurately calculate the longitude and latitude of the target, thereby meeting the actual needs of precise positioning.
[0087] In large-scale, multi-target monitoring scenarios, two obstacle detectors are used to work together to cover the entire area. After the operator manually calibrates the target in the main obstacle detector, the calibration data is sent to the secondary obstacle detector. The secondary obstacle detector performs a secondary cross-confirmation of the manually calibrated target during idle time when not performing tasks to ensure the accuracy of recognition and positioning. After the secondary confirmation, the final target recognition result is fed back to the model training module to promptly correct the obstacle detection algorithm so that the model gradually adapts to the new environment or target characteristics. At the same time, by accumulating more manual calibration data, the system can continuously update the model and improve its ability to accurately identify and locate multiple targets in complex scenarios.
[0088] To summarize, this application significantly improves the accuracy, flexibility and adaptability of the target recognition and detection system by enhancing the manual calibration function, adopting a multi-obstacle detector collaboration mechanism, and realizing dynamic adjustment and adaptive optimization of the model. It can meet the needs of wide-area monitoring and multi-target accurate recognition and positioning in complex scenarios, and overcomes the shortcomings of the existing technology in target recognition accuracy, coverage range and system sustainable optimization capabilities. It can achieve more accurate and reliable target recognition in complex and changeable underwater environments by combining the operator's intuitive judgment and the machine's automatic processing.
[0089] See also Figure 4 Corresponding to the above-mentioned embodiment of the multi-obstacle detector underwater target recognition method based on manual calibration, the embodiment of the present application provides a multi-obstacle detector underwater target recognition device based on manual calibration, the device comprising:
[0090] The data acquisition module 1001 is used to receive and decode the real-time video streams of multiple obstacle detectors and display them in the form of images;
[0091] The calibration module 1002 is used to obtain an external frame selection input for an unrecognized target, and calibrate the unrecognized target through the frame selection input; at the same time, obtain an external target type T1 and confidence C1 input;
[0092] The latitude and longitude calculation module 1003 is used to calculate the longitude and latitude of the calibration target based on the center point of the box selection input;
[0093] The secondary cross-confirmation module 1004 is used to send the latitude and longitude of the calibration target, the target type T1 and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system;
[0094] The recognition model updating module 1005 is used to continuously send the updated recognition results to the model training modules of the multiple obstacle detectors to update the recognition model.
[0095] Furthermore, the latitude and longitude calculation module 1003 is specifically used for:
[0096] Based on the known obstacle detector coordinate system, calculating the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system;
[0097] Based on the known north angle of the obstacle detector, converting the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system into the distance relative to the obstacle detector in the UTM coordinate system;
[0098] Based on the UTM coordinate system projection method, the known longitude and latitude of the obstacle detector are converted into coordinates in the UTM coordinate system;
[0099] According to the coordinates of the obstacle detector in the UTM coordinate system, the distance of the calibration target relative to the obstacle detector is superimposed in the X direction and the Y direction to obtain the coordinates of the calibration target in the UTM coordinate system;
[0100] According to the coordinates of the calibration target in the UTM coordinate system, the coordinate values of the calibration target in the UTM coordinate system are converted into longitude and latitude using the UTM coordinate system projection method.
[0101] Furthermore, the secondary cross-confirmation module 1004 is specifically used for:
[0102] The obstacle detector is configured to perform secondary recognition on the calibration target according to the acquired latitude and longitude of the calibration target, and to provide the target type T2 and confidence C2 according to the built-in recognition algorithm;
[0103] According to the comparison, when the target type T1 and the target type T2 are consistent, the target type is determined to be T1, and the final confidence is (C1+C2) / 2;
[0104] When the target type T1 and the target type T2 are inconsistent, the two detection results are saved locally, and wait for external input to be obtained to make a final determination and update the recognition result.
[0105] Furthermore, the secondary cross-confirmation module 1004 is also used for:
[0106] If the obstacle detector cannot identify the calibration target during the secondary identification, the recognition result of the system is updated according to the manual calibration data input externally.
[0107] The above-mentioned multi-obstacle detector underwater target recognition device based on manual calibration implements the steps and various processes of the above-mentioned multi-obstacle detector underwater target recognition method based on manual calibration, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
[0108] See also Figure 5 Corresponding to the above-mentioned embodiment of the multi-obstacle detector underwater target recognition method based on manual calibration, an embodiment of the present application provides an electronic device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, the steps and various processes of the above-mentioned embodiment of the multi-obstacle detector underwater target recognition method based on manual calibration are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
[0109] The memory 1009 can be used to store software programs and various data. The memory 1009 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instructions required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 1009 may include a volatile memory or a non-volatile memory, or the memory 1009 may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM). The memory 1009 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.
[0110] The processor 1010 may include one or more processing units; optionally, the processor 1010 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 1010.
[0111] Corresponding to the above-mentioned embodiment of the multi-obstacle detector underwater target identification method based on manual calibration, the embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by the processor, the steps and various processes of the above-mentioned embodiment of the multi-obstacle detector underwater target identification method based on manual calibration are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
[0112] The processor is the processor in the electronic device described in the above embodiment of the present application. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.
[0113] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.
[0114] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.
[0115] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.
Claims
1. A multi-obstacle detector underwater target recognition method based on manual calibration, characterized in that: include: Receive and decode the real-time video streams of multiple obstacle detectors and display them in the form of images; For an unrecognized target, an external frame selection input is obtained, and the unrecognized target is calibrated by the frame selection input; at the same time, an external target type T1 and confidence C1 input are obtained; Calculate the longitude and latitude of the calibration target based on the center point of the frame selection input; Sending the latitude and longitude of the calibration target, the target type T1 and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system; The updated recognition results are continuously sent to the model training modules of the plurality of obstacle detectors to update the recognition model.
2. The method for underwater target recognition using multiple obstacle detectors based on manual calibration according to claim 1, characterized in that: The calculating the longitude and latitude of the calibration target based on the center point of the frame selection input includes: Based on the known obstacle detector coordinate system, calculating the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system; Based on the known north angle of the obstacle detector, converting the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system into the distance relative to the obstacle detector in the UTM coordinate system; Based on the UTM coordinate system projection method, the known longitude and latitude of the obstacle detector are converted into coordinates in the UTM coordinate system; According to the coordinates of the obstacle detector in the UTM coordinate system, the distance of the calibration target relative to the obstacle detector is superimposed in the X direction and the Y direction to obtain the coordinates of the calibration target in the UTM coordinate system; According to the coordinates of the calibration target in the UTM coordinate system, the coordinate values of the calibration target in the UTM coordinate system are converted into longitude and latitude using the UTM coordinate system projection method.
3. The method for underwater target recognition using multiple obstacle detectors based on manual calibration according to claim 1, characterized in that: The sending of the latitude and longitude of the calibration target, the target type T1 and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system, includes: The obstacle detector is configured to perform secondary recognition on the calibration target according to the acquired latitude and longitude of the calibration target, and to provide the target type T2 and confidence C2 according to the built-in recognition algorithm; According to the comparison, when the target type T1 and the target type T2 are consistent, the target type is determined to be T1, and the final confidence is (C1+C2) / 2; When the target type T1 and the target type T2 are inconsistent, the two detection results are saved locally, and wait for external input to be obtained to make a final determination and update the recognition result.
4. The method for underwater target recognition using multiple obstacle detectors based on manual calibration according to claim 3 is characterized in that: If the obstacle detector cannot identify the calibration target during the secondary identification, the recognition result of the system is updated according to the manual calibration data input externally.
5. A multi-obstacle detector underwater target recognition device based on manual calibration, characterized in that: include: A data acquisition module is used to receive and decode the real-time video streams of multiple obstacle detectors and display them in the form of images; A calibration module is used to obtain an external frame selection input for an unrecognized target, and calibrate the unrecognized target through the frame selection input; at the same time, obtain an external target type T1 and confidence C1 input; A latitude and longitude calculation module, used to calculate the longitude and latitude of the calibration target based on the center point of the box selection input; A secondary cross-confirmation module is used to send the latitude and longitude of the calibration target, the target type T1 and the confidence C1 to one of the obstacle detectors in a non-detection task state to perform a secondary cross-confirmation on the calibration target, obtain a secondary recognition result and update the recognition result of the system; The recognition model updating module is used to continuously send the updated recognition results to the model training modules of the multiple obstacle detectors to update the recognition model.
6. The multi-obstacle detector underwater target recognition device based on manual calibration according to claim 5 is characterized in that: The latitude and longitude calculation module is specifically used for: Based on the known obstacle detector coordinate system, calculating the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system; Based on the known north angle of the obstacle detector, converting the distance of the calibration target relative to the obstacle detector in the obstacle detector coordinate system into the distance relative to the obstacle detector in the UTM coordinate system; Based on the UTM coordinate system projection method, the known longitude and latitude of the obstacle detector are converted into coordinates in the UTM coordinate system; According to the coordinates of the obstacle detector in the UTM coordinate system, the distance of the calibration target relative to the obstacle detector is superimposed in the X direction and the Y direction to obtain the coordinates of the calibration target in the UTM coordinate system; According to the coordinates of the calibration target in the UTM coordinate system, the coordinate values of the calibration target in the UTM coordinate system are converted into longitude and latitude using the UTM coordinate system projection method.
7. The multi-obstacle detector underwater target recognition device based on manual calibration according to claim 5, characterized in that: The secondary cross-confirmation module is specifically used for: The obstacle detector is configured to perform secondary recognition on the calibration target according to the acquired latitude and longitude of the calibration target, and to provide the target type T2 and confidence C2 according to the built-in recognition algorithm; According to the comparison, when the target type T1 and the target type T2 are consistent, the target type is determined to be T1, and the final confidence is (C1+C2) / 2; When the target type T1 and the target type T2 are inconsistent, the two detection results are saved locally, and wait for external input to be obtained to make a final determination and update the recognition result.
8. The multi-obstacle detector underwater target recognition device based on manual calibration according to claim 7 is characterized in that: The secondary cross-confirmation module is further used for: If the obstacle detector cannot identify the calibration target during the secondary identification, the recognition result of the system is updated according to the manual calibration data input externally.
9. An electronic device, characterized in that: The electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the multi-obstacle detector underwater target recognition method based on manual calibration as described in any one of claims 1 to 4 are implemented.
10. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of the multi-obstacle detector underwater target recognition method based on manual calibration as described in any one of claims 1 to 4 are implemented.