Method and device for evaluating target recognition capability

Abstract

The invention provides a target recognition capability evaluation method and device. The method comprises the following steps: acquiring a sea surface image; determining an image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; and comparing the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluating the target classification algorithm model according to the comparison result. And a target classification algorithm model is constructed to simulate the autonomous cognitive ability (water surface target recognition ability) of the unmanned system. According to requirements, the system has the capability of classifying ships, civilian ships and buoys. Secondly, constructing a test sample according to polymorphic behavior exploration research results; and finally, testing the capability of the constructed model according to a test process, and calculating an evaluation index according to a result, thereby improving the evaluation efficiency and accuracy of the environment perception capability of the virtual visual target of the unmanned ship.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned surface systems, and particularly relates to a method and device for evaluating the target recognition ability. Background Art

[0002] At present, an environmental perception test device generates a series of virtual visual scene targets and sends these virtual visual scenes to an unmanned boat at a fixed frequency; the unmanned boat identifies the virtual visual scene targets and feeds back the recognition results to the test device, and at the same time, the command and control center dynamically displays the visual scene; the environmental perception test device compares the recognition results fed back by the unmanned boat with the real results, and statistics the recognition accuracy rate and timeliness rate of virtual targets. Due to the influence of the environment, the current environmental perception test device cannot accurately and timely evaluate the environmental perception ability of the virtual visual scene targets of the unmanned boat. Therefore, how to improve the evaluation method of the environmental perception ability of the virtual visual scene targets of the unmanned boat has become a technical problem that needs to be solved urgently in this field. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and device for evaluating the target recognition ability.

[0004] According to the first aspect of the present invention, there is provided an evaluation of the target recognition ability, including,

[0005] Obtaining a sea surface image;

[0006] Determining an image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network with sample frame images;

[0007] Comparing the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluating the target classification algorithm model according to the comparison result.

[0008] Optionally, the test sample is obtained by processing an initial sea surface image according to the changes in the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture.

[0009] Optionally, the method further includes:

[0010] Setting corresponding value ranges for the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture according to different sea state levels and fog levels;

[0011] Grouping the value ranges of the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture to obtain a grouping result when the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture change;

[0012] Evaluate the target classification algorithm model according to the grouping result.

[0013] Optionally, when the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture change, group the value ranges of the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture to obtain a grouping result, including:

[0014] Obtain the grouping result according to the natural disturbance transformation corresponding to different sea conditions, where the natural disturbance transformation corresponding to different sea conditions at least includes: rotation, scaling, perspective, positive barrel distortion, negative barrel distortion, positive pillow distortion, negative pillow distortion, positive tangential distortion, negative tangential distortion, Gaussian blur, angular motion blur, and velocity motion blur.

[0015] Optionally, when the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture change, group the value ranges of the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture to obtain a grouping result, including:

[0016] Obtain the grouping result according to the natural disturbance transformation corresponding to different fogs, where the natural disturbance transformation corresponding to different fogs at least includes: mean blur, contrast, brightness, and saturation.

[0017] According to a second aspect of the present invention, there is provided an evaluation device for target recognition ability, including the evaluation method for target recognition ability according to any one of the first aspects of the present invention, including:

[0018] An acquisition module for acquiring a sea surface image;

[0019] A determination module for determining an image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network with sample frame images.

[0020] An evaluation module for comparing the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluating the target classification algorithm model according to the comparison result.

[0021] Optionally, the test sample is obtained by processing an initial sea surface image according to the change conditions of the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture.

[0022] Optionally, the evaluation module is used for:

[0023] Set corresponding value ranges for the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture according to different sea state levels and fog levels;

[0024] When the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture change, group the value ranges of the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture to obtain a grouping result;

[0025] Evaluate the target classification algorithm model according to the grouping result.

[0026] Optionally, the evaluation module is used for:

[0027] Obtain the grouping result according to the natural disturbance transformation corresponding to different sea states, where the natural disturbance transformation corresponding to different sea states at least includes: rotation, scaling, perspective, positive barrel distortion, negative barrel distortion, positive pillow distortion, negative pillow distortion, positive tangential distortion, negative tangential distortion, Gaussian blur, angular motion blur, and velocity motion blur.

[0028] Optionally, the evaluation module is used for:

[0029] Obtain the grouping result according to the natural disturbance transformation corresponding to different fogs, where the natural disturbance transformation corresponding to different fogs at least includes: mean blur, contrast, brightness, and saturation.

[0030] In a third aspect, the present application shows an electronic device, which includes: a processor; a memory for storing processor-executable instructions; wherein, the processor is configured to execute the method described in any of the above aspects.

[0031] In a fourth aspect, the present application shows a non-transitory computer-readable storage medium, when the instructions in the storage medium are executed by a processor of an electronic device, enabling the electronic device to execute the method described in any of the above aspects.

[0032] In a fifth aspect, the present application shows a computer program product, when the instructions in the computer program product are executed by a processor of an electronic device, enabling the electronic device to execute the method described in any of the above aspects.

[0033] The beneficial effects brought by the present invention are as follows:

[0034] As can be seen from the above solution, the embodiments of the present invention provide a method and device for evaluating the target recognition ability, including: obtaining a sea surface image; determining an image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network with sample frame images; comparing the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluating the target classification algorithm model according to the comparison result. A target classification algorithm model will be constructed to simulate the autonomous cognitive ability (water surface target recognition ability) of an unmanned system. According to the requirements, it is necessary to have the classification ability of ships, civilian ships and buoys. Secondly, test samples are constructed according to the research results of polymorphic behavior exploration. Finally, the ability of the constructed model is tested according to the test process, and evaluation indexes are calculated according to the results, so as to improve the evaluation efficiency and accuracy of the environmental perception ability of the virtual vision target of the unmanned boat. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 FIG. is a schematic flowchart of a method for evaluating the target recognition ability provided by an embodiment;

[0036] Figure 2 FIG. is a schematic diagram of sample generation under rotation transformation provided by an embodiment;

[0037] Figure 3 FIG. is a schematic diagram of sample generation under scaling transformation provided by an embodiment;

[0038] Figure 4 FIG. is a schematic diagram of sample generation under perspective transformation provided by an embodiment;

[0039] Figure 5 FIG. is a schematic diagram of sample generation under positive barrel distortion provided by an embodiment;

[0040] Figure 6 FIG. is a schematic diagram of sample generation under negative barrel distortion provided by an embodiment;

[0041] Figure 7 FIG. is a schematic diagram of sample generation under positive pincushion distortion provided by an embodiment;

[0042] Figure 8 FIG. is a schematic diagram of sample generation under negative pincushion distortion provided by an embodiment;

[0043] Figure 9 FIG. is a schematic diagram of sample generation under positive tangential distortion provided by an embodiment;

[0044] Figure 10 FIG. is a schematic diagram of sample generation under negative tangential distortion provided by an embodiment;

[0045] Figure 11 Schematic diagram of sample generation under Gaussian blur transformation provided according to an embodiment;

[0046] Figure 12 Schematic diagram of sample generation under motion blur (angle change) provided according to an embodiment;

[0047] Figure 13 Schematic diagram of sample generation under motion blur (speed change) transformation provided according to an embodiment;

[0048] Figure 14 Schematic diagram of sample generation under mean blur transformation provided according to an embodiment;

[0049] Figure 15 Schematic diagram of sample generation under contrast transformation provided according to an embodiment;

[0050] Figure 16 Schematic diagram of sample generation under brightness change provided according to an embodiment;

[0051] Figure 17 Schematic diagram of sample generation under saturation change provided according to an embodiment;

[0052] Figure 18 Schematic diagram of a sample example of a water surface target recognition dataset provided according to an embodiment;

[0053] Figure 19 Schematic diagram of a sample example of rotation transformation data provided according to an embodiment;

[0054] Figure 20 Schematic diagram of a sample example of scale transformation data provided according to an embodiment;

[0055] Figure 21 Block diagram of an evaluation device for the target recognition ability of the present application;

[0056] Figure 22 Block diagram of an electronic device of the present application;

[0057] Figure 23 Block diagram of a computer-readable storage medium of the present application. Specific embodiments

[0058] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0059] Refer to Figure 1 , a step flowchart of an evaluation method for a target recognition ability of the present application is shown. This method can be applied to an electronic device. Specifically, the method may include the following steps:

[0060] S101. Obtain a sea surface image;

[0061] S102. Determine an image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network with sample frame images;

[0062] S103. Compare the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluate the target classification algorithm model according to the comparison result.

[0063] First, a target classification algorithm model will be constructed to simulate the autonomous cognitive ability (water surface target recognition ability) of an unmanned system. According to requirements, it is necessary to have the classification ability for ships, civilian ships, and buoys. Secondly, test samples are constructed based on the research results of polymorphic behavior exploration. Finally, the ability of the constructed model is tested according to the test process, and evaluation indexes are calculated according to the results. The target recognition test and evaluation ability has been possessed, and the coverage rate of the cognitive ability test is not less than 85%.

[0064] Another embodiment of the present application further supplements and explains the evaluation method for the target recognition ability provided in the above embodiment.

[0065] Optionally, the test sample is obtained by processing an initial sea surface image according to the changes in the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture.

[0066] Specifically, in combination with the task background of the water surface unmanned system, the change of sea conditions will lead to the change of imaging key factors such as the illumination condition, pose, perspective distortion, visibility, and motion blur of the picture, so as to simulate different sea condition scenarios. As shown in Table 1, they are the key influencing factors for image target recognition.

[0067] Table 1

[0068]

[0069] For different sea condition levels and fog levels, the above-transformed parameters have a definite value range. To better describe the performance change of the model in different environments, it is necessary to group the value range of the transformed parameters and evaluate the performance of the model under different groups of parameters respectively. First, the environmental factors are classified, and the results are shown in Table 2:

[0070] Table 2

[0071]

[0072]

[0073] Optionally, the method further includes:

[0074] Setting corresponding value ranges for the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur according to different sea state levels and fog levels;

[0075] Grouping the value ranges of the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur when the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur change, to obtain a grouping result;

[0076] Evaluating the target classification algorithm model according to the grouping result.

[0077] Optionally, when the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur change, grouping the value ranges of the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur to obtain a grouping result, including:

[0078] Obtaining the grouping result according to the natural disturbance transformations corresponding to different sea states, where the natural disturbance transformations corresponding to different sea states at least include: rotation, scaling, view angle, positive barrel distortion, negative barrel distortion, positive pincushion distortion, negative pincushion distortion, positive tangential distortion, negative tangential distortion, Gaussian blur, angular motion blur, and velocity motion blur.

[0079] Optionally, when the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur change, grouping the value ranges of the picture lighting condition, pose situation, perspective distortion, visibility, and motion blur to obtain a grouping result, including:

[0080] Obtaining the grouping result according to the natural disturbance transformations corresponding to different fogs, where the natural disturbance transformations corresponding to different fogs at least include: mean blur, contrast, brightness, and saturation.

[0081] By qualitatively comparing the videos of the actual ship traveling under different environmental levels with the transformed pictures, the parameter grouping results are obtained, as shown in Table 3 and Table 4. Table 3 shows the natural disturbance factors and classifications corresponding to different sea states, and Table 4 shows the natural disturbance factors and classifications corresponding to different fogs:

[0082] Table 3

[0083]

[0084]

[0085] Table 4

[0086]

[0087] The generation of test samples for target recognition ability is as follows:

[0088] 1) Generation of rotation transformation test samples

[0089] The jolting of the intelligent system platform causes a change in the relative position between it and the object, resulting in image rotation, which will have a certain impact on the operation result. It is planned to generate test data from 12 sampling points with an interval of 300 as a sampling point. As Figure 2 shown.

[0090]

[0091] 2) Generation of scaling transformation test samples

[0092] The distance from the object poses a great challenge to the working condition of the intelligent system platform in actual deployment. Exploring the limit distance at which the model can work properly is a good way to understand the failure boundary of the model. It is planned to generate test data from 9 sampling points starting from an object full of images, which are 9 / 10, 8 / 10,..., 1 / 10. As Figure 3 shown.

[0093]

[0094] 3) Generation of perspective transformation test samples

[0095] It is planned to generate test data from 10 sampling points of the observation perspective. The transformation formulas under different perspectives are as shown, as Figure 4 shown.

[0096]

[0097] Generation of distortion transformation test samples

[0098] Due to the compositional defects of the imaging system and the perturbation of the optical medium, the occurrence of distortion almost accompanies all data in real working conditions. Since the characteristic changes brought by it will have a certain impact on the results. It is planned to include a total of six sub-items of radial distortion and tangential distortion, with 10 sampling points for each item to generate a total of 60 groups of test data.

[0099] 4) Positive barrel distortion, as Figure 5 shown;

[0100] x′ = x(1 + k 1 r 2 + k 2 r 4 + k3 r 6 )

[0101] y′ = y(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0102] r 2 = x 2 + y 2

[0103] k 1 > 0, k 2 > 0

[0104] 5) Negative barrel distortion, such as Figure 6 shown.

[0105] x′ = x(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0106] y′ = y(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0107] r 2 = x 2 + y 2

[0108] k 1 > 0, k 2 < 0

[0109] 6) Positive pincushion distortion, such as Figure 7 shown;

[0110] x′ = x(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0111] y′ = y(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0112] r 2 = x 2 + y 2

[0113] k 1 < 0, k 2 > 0

[0114] 7) Negative pincushion distortion, such as Figure 8 shown;

[0115] x' = x(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0116] y' = y(1 + k 1 r 2 + k 2 r 4 + k 3 r 6 )

[0117] r 2 = x 2 + y 2

[0118] k 1 < 0, k 2 < 0

[0119] 8) Positive tangential distortion, such as Figure 9 shown;

[0120] x' = x + [2p 1 xy + p 2 (r 2 + 2x 2 )]

[0121] y' = y + [2p 2 xy + p 1 (r 2 + 2y 2 )]

[0122] r 2 = x 2 + y 2

[0123] p 1 > 0, p 2 > 0

[0124] 9) Negative tangential distortion, such as Figure 10 shown;

[0125] x' = x + [2p1 xy + p 2 (r 2 + 2x 2 )]

[0126] y′ = y + [2p 2 xy + p 1 (r 2 + 2y 2 )]

[0127] r 2 = x 2 + y 2

[0128] p 1 <0, p 2 <0

[0129] Fuzzy transformation test sample generation. Since defocusing in the intelligent system platform and imaging system will result in data degradation, which will harm the intelligent system. It is planned to select 10 sampling points for each of the four test items of Gaussian blur, defocus blur, and motion blur to generate 40 groups of test data.

[0130] (1) Gaussian blur, as Figure 11 shown,

[0131]

[0132] (2) Motion blur (angle change) as Figure 12 shown,

[0133]

[0134] (3) Motion blur (speed change), as Figure 13 ;

[0135]

[0136] (4) Mean blur, as Figure 14 shown,

[0137]

[0138] Lighting transformation test sample generation. In the actual working condition, the lighting situation is complex and changeable, which will cause a large pattern difference between the collected data and the training data. It is planned to generate three test items including brightness, saturation, and contrast, and take 10 sampling points for each item to generate a total of 30 groups of test data.

[0139] (1) Contrast, as Figure 15 shown.

[0140] s i = T i (ri ), where \(i = 1, 2, \ldots, n\)

[0141] (2) Brightness, as shown in Figure 16 .

[0142] s i = T i (r i ), where \(i = 1, 2, \ldots, n\)

[0143] (3) Saturation, as shown in Figure 17 .

[0144] s i = T i (r i ), where \(i = 1, 2, \ldots, n\)

[0145] Construct an image-based water surface target classification model to simulate the autonomous cognitive ability of an unmanned surface system. This model can distinguish 11 types of targets including ships, civilian boats, and buoys. The information of the training and test data sets used is as follows. See the example pictures in Figure 18 .

[0146] ① Training data set: Capacity 5500, 11 classes in total, 500 pictures for each class;

[0147] ② Test data set: Capacity 1100, 11 classes in total, 100 pictures for each class;

[0148] Select the widely used Res-Net 18 convolutional neural network (the network structure is shown in the following figure) as the test object for the classifier, train the recognition model using the training data set, and test the test data set using the model. It is found that the accuracy rate of the model on the test set reaches 84.55%.

[0149] For the purpose of exploring the failure boundary of the model, the test data respectively simulate the natural environment disturbances and human attacks that may be encountered after the deployment of the intelligent system.

[0150] ① Natural disturbances include 16 influencing factors such as rotation, cropping, distortion, blur, hue, saturation, and brightness. An average of 10 sampling points are taken for each influencing factor to construct a test profile.

[0151] ② Human attacks use 16 attack methods including black box / white box, single-step / iterative, targeted / untargeted, \(L_{\infty} / L_0\), etc. An average of 10 sampling points are taken for each influencing factor to construct a test profile.

[0152] For each original sample in the original image data set, perform a rotation transformation, and the rotation angles are respectively set to:

[0153] [4°, 8°, 12°, 16°, 20°, 24°, 28°, 32°, 36°]

[0154] The transformed sample is shown as Figure 19 follows.

[0155] For each original sample in the original image dataset, perform a scaling transformation, with the scaling factor Scale (the s value in the transformation formula):

[0156] [0.05, 0.10, 0.15, 0.20, 0.25, 0.30, 0.35, 0.40, 0.45]

[0157] The sample after the scale transformation is shown as Figure 20 follows.

[0158] The embodiment of the present invention provides a method and device for evaluating the target recognition ability, including: obtaining a sea surface image; determining the image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training the Res-Net 18 convolutional neural network with sample frame images; comparing the sea surface image with the test samples corresponding to the image type to obtain a comparison result, and evaluating the target classification algorithm model according to the comparison result. A target classification algorithm model will be constructed to simulate the autonomous cognitive ability (water surface target recognition ability) of the unmanned system. According to the requirements, it is necessary to have the classification ability of ships, civilian ships and buoys. Secondly, test samples are constructed according to the research results of polymorphic behavior exploration. Finally, the ability of the constructed model is tested according to the test process, and evaluation indicators are calculated according to the results to improve the evaluation efficiency and accuracy of the environmental perception ability of the virtual vision target of the unmanned boat.

[0159] It should be noted that each implementable manner in this embodiment can be implemented alone, or can be combined in any combination manner without conflict. The present application makes no limitation.

[0160] Another embodiment of the present application provides an apparatus for evaluating the target recognition ability, which is used to execute the method for evaluating the target recognition ability provided in the above embodiment.

[0161] As Figure 21 shown, it is a schematic structural diagram of the apparatus for evaluating the target recognition ability provided in the embodiment of the present application. The apparatus includes an acquisition module 2101, a determination module 2102 and an evaluation module 2103, where:

[0162] The acquisition module 2101 is used to obtain a sea surface image;

[0163] The determination module 2102 is configured to determine the image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network with sample frame images;

[0164] The evaluation module 2103 is configured to compare the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluate the target classification algorithm model according to the comparison result.

[0165] Regarding the device in this embodiment, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0166] Another embodiment of the present application further supplements the evaluation device for the target recognition ability provided in the above embodiment.

[0167] Optionally, the test sample is obtained by processing an initial sea surface image according to changes in picture illumination conditions, pose conditions, perspective distortion, visibility, and motion blur.

[0168] Optionally, the evaluation module is configured to:

[0169] Set corresponding value ranges for picture illumination conditions, pose conditions, perspective distortion, visibility, and motion blur according to different sea state levels and fog levels;

[0170] Group the value ranges of picture illumination conditions, pose conditions, perspective distortion, visibility, and motion blur when the picture illumination conditions, pose conditions, perspective distortion, visibility, and motion blur change to obtain a grouping result;

[0171] Evaluate the target classification algorithm model according to the grouping result.

[0172] Optionally, the evaluation module is configured to:

[0173] Obtain a grouping result according to natural perturbation transformations corresponding to different sea states, where the natural perturbation transformations corresponding to different sea states at least include: rotation, scaling, perspective, positive barrel distortion, negative barrel distortion, positive pincushion distortion, negative pincushion distortion, positive tangential distortion, negative tangential distortion, Gaussian blur, angular motion blur, and velocity motion blur.

[0174] Optionally, the evaluation module is configured to:

[0175] Obtain a grouping result according to natural perturbation transformations corresponding to different fogs, where the natural perturbation transformations corresponding to different fogs at least include: mean blur, contrast, brightness, and saturation.

[0176] The embodiments of the present invention provide a method and apparatus for evaluating the target recognition ability, including: obtaining a sea surface image; determining an image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network with sample frame images; comparing the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluating the target classification algorithm model according to the comparison result. A target classification algorithm model will be constructed to simulate the autonomous cognitive ability (water surface target recognition ability) of an unmanned system. According to requirements, it is necessary to have the classification ability of ships, civilian ships, and buoys. Secondly, test samples are constructed according to the research results of polymorphic behavior exploration. Finally, the ability of the constructed model is tested according to the test process, and evaluation indexes are calculated according to the results to improve the evaluation efficiency and accuracy of the environmental perception ability of the virtual vision target of the unmanned boat.

[0177] For the apparatus embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for the relevant parts, please refer to the partial description of the method embodiments.

[0178] Optionally, the embodiments of the present application further provide an electronic device, including: a processor, a memory, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, it implements each process of the above method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described here again.

[0179] The embodiments of the present application further provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, it implements each process of the above method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described here again. Among them, the computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.

[0180] Figure 22 It is a block diagram of an electronic device 800 shown in the present application. For example, the electronic device 800 can be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0181] Refer to Figure 22 , the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, a sensor component 814, and a communication component 816.

[0182] The processing component 802 generally controls the overall operation of the electronic device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above methods. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0183] The memory 804 is configured to store various types of data to support the operation of the device 800. Examples of such data include instructions for any application or method operating on the electronic device 800, contact data, phone book data, messages, images, videos, etc. The memory 804 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disks, or optical disks.

[0184] The power component 806 provides power to various components of the electronic device 800. The power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the electronic device 800.

[0185] The multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may not only sense the boundaries of the touch or swipe actions, but also detect the duration and pressure associated with the touch or swipe operation. In some embodiments, the multimedia component 808 includes a front camera and / or a rear camera. When the device 800 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each of the front camera and the rear camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0186] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when the electronic device 800 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal can be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, the audio component 810 further includes a speaker for outputting audio signals.

[0187] The I / O interface 812 provides an interface between the processing component 802 and peripheral interface modules, and the peripheral interface modules may be a keyboard, a click wheel, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a power button, and a lock button.

[0188] The sensor component 814 includes one or more sensors for providing an assessment of the state of various aspects of the electronic device 800. For example, the sensor component 814 can detect the on / off state of the device 800, the relative positioning of components, such as the display and keypad of the electronic device 800. The sensor component 814 can also detect a change in the position of the electronic device 800 or a component of the electronic device 800, the presence or absence of user contact with the electronic device 800, the orientation or acceleration / deceleration of the electronic device 800, and the temperature change of the electronic device 800. The sensor component 814 can include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor component 814 can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 814 can further include an acceleration sensor, a gyro sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0189] The communication component 816 is configured to facilitate communication between the electronic device 800 and other devices in a wired or wireless manner. The electronic device 800 can access a wireless network based on communication standards, such as WiFi, a carrier network (such as 2G, 3G, 4G, or 5G), or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast operation information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0190] In an exemplary embodiment, the electronic device 800 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the above method.

[0191] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 804 including instructions, and the above instructions can be executed by a processor 820 of the electronic device 800 to complete the above method. For example, the non-transitory computer-readable storage medium may be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0192] Figure 23 FIG. is a block diagram of a computer-readable storage medium 1900 shown in the present application. For example, the computer-readable storage medium 1900 may be provided as a server.

[0193] Refer to Figure 23 , the computer-readable storage medium 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by a memory 1932 for storing instructions executable by the processing component 1922, such as application programs. The application programs stored in the memory 1932 may include one or more modules each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute instructions to perform the above method.

[0194] The computer-readable storage medium 1900 may further include a power component 1926 configured to perform power management of the computer-readable storage medium 1900, a wired or wireless network interface 1950 configured to connect the computer-readable storage medium 1900 to a network, and an input / output (I / O) interface 1958. The computer-readable storage medium 1900 may operate based on an operating system stored in the memory 1932, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or the like.

[0195] It should be noted that in this article, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including such element.

[0196] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described example methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc) and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of this application.

[0197] The embodiments of this application have been described above in conjunction with the accompanying drawings. However, this application is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of this application, those of ordinary skill in the art can also make many forms without departing from the purpose of this application and the scope protected by the claims, and all of them fall within the protection scope of this application.

[0198] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the embodiments of this application can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0199] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0200] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of the devices or units can be in electrical, mechanical or other forms.

[0201] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0202] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0203] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable 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 each embodiment of the present application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0204] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0205] The above is the preferred implementation manner of the present invention. It should be noted that for those of ordinary skill in the art in this technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for evaluating target recognition capability, characterized in that: The method comprises: Acquire sea surface images; Determining the image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network using sample screen images; The sea surface image is compared with a test sample corresponding to the image type to obtain a comparison result, and the target classification algorithm model is evaluated according to the comparison result.

2. The target recognition capability evaluation method according to claim 1, characterized in that: The test sample is obtained by processing the initial sea surface image according to changes in the image lighting conditions, posture, perspective distortion, visibility and dynamic blur.

3. The target recognition capability evaluation method according to claim 2, characterized in that: The method further comprises: According to different sea conditions and fog levels, corresponding value ranges are set for the lighting conditions, posture conditions, perspective distortion, visibility and dynamic blur of the image; When the illumination condition, posture, perspective distortion, visibility and dynamic blur of the picture change, grouping the value ranges of the illumination condition, posture, perspective distortion, visibility and dynamic blur of the picture to obtain a grouping result; The target classification algorithm model is evaluated according to the grouping results.

4. The target recognition capability evaluation method according to claim 3, characterized in that: When the illumination condition, posture condition, perspective distortion, visibility and dynamic blur of the picture change, the value ranges of the illumination condition, posture condition, perspective distortion, visibility and dynamic blur of the picture are grouped to obtain grouping results, including: The grouping result is obtained according to the natural disturbance transformation corresponding to different sea conditions, wherein the natural disturbance transformation corresponding to different sea conditions at least includes: rotation, scaling, viewing angle, positive barrel distortion, negative barrel distortion, positive pincushion distortion, negative pincushion distortion, positive tangential distortion, negative tangential distortion, Gaussian blur, angular motion blur and velocity motion blur.

5. The target recognition capability evaluation method according to claim 3, characterized in that: When the illumination condition, posture condition, perspective distortion, visibility and dynamic blur of the picture change, the value ranges of the illumination condition, posture condition, perspective distortion, visibility and dynamic blur of the picture are grouped to obtain grouping results, including: The grouping result is obtained according to the natural disturbance transformation corresponding to different fogs, wherein the natural disturbance transformation corresponding to the different fogs at least includes: mean blur, contrast, brightness and saturation.

6. A target recognition capability evaluation device, characterized in that: The device comprises: An acquisition module, used for acquiring sea surface images; A determination module, used to determine the image type corresponding to the sea surface image according to the sea surface image and a pre-trained target classification algorithm model; the target classification algorithm model is obtained by training a Res-Net 18 convolutional neural network using sample screen images; An evaluation module is used to compare the sea surface image with a test sample corresponding to the image type to obtain a comparison result, and evaluate the target classification algorithm model according to the comparison result.

7. The target recognition capability evaluation device according to claim 6, characterized in that: The test sample is obtained by processing the initial sea surface image according to changes in the image lighting conditions, posture, perspective distortion, visibility and dynamic blur.

8. The target recognition capability evaluation device according to claim 7, characterized in that: The evaluation module is used to: According to different sea conditions and fog levels, corresponding value ranges are set for the lighting conditions, posture conditions, perspective distortion, visibility and dynamic blur of the image; When the illumination condition, posture, perspective distortion, visibility and dynamic blur of the picture change, grouping the value ranges of the illumination condition, posture, perspective distortion, visibility and dynamic blur of the picture to obtain a grouping result; The target classification algorithm model is evaluated according to the grouping results.

9. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the method according to any one of claims 1 to 5 when executed by the processor.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images