Active light projection camouflage method based on EOT conversion, electronic equipment and medium

Abstract

The invention discloses an active light projection camouflage method based on EOT transformation, electronic equipment and a medium. The method comprises the following steps: constructing an adaptive color transformation EOT algorithm; configuring an adaptive projection compensation strategy fed back by ambient light; based on a self-adaptive projection compensation strategy fed back by ambient light, processing the texture pattern through a self-adaptive color transformation (EOT) algorithm to obtain a projection pattern; a plurality of sampling points are selected on the surface of the target object, and the position of each sampling point is detected in real time, so that the pose of the target object is estimated; and projecting the projection pattern to the surface of the target object according to the pose of the target object so as to realize camouflage.

Description

Technical Field

[0001] This invention belongs to the field of computer vision and image processing, and particularly relates to an active optical projection camouflage method, electronic device, and medium based on EOT transformation. Background Technology

[0002] The basic idea of ​​active light projection camouflage is to use optical devices such as projectors to project carefully designed lighting patterns onto the surface of a target object, thereby disrupting the object's visual characteristics and causing the visual system (such as a camera) to make incorrect perceptions and judgments.

[0003] Despite the impressive performance of active light projection camouflage, several challenges remain in practical applications. First, there's the deformation issue: when the target object undergoes non-rigid deformations such as bending or folding, the pre-designed projection pattern becomes distorted, reducing its camouflage effectiveness. Second, ambient lighting plays a role; the projection pattern superimposed on changing ambient light makes the visual features captured by the sensing system unstable. Furthermore, the movement of the target object itself introduces disturbances, causing the projection area to shift. These factors all contribute to reducing the effectiveness of projection camouflage to varying degrees. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an active optical projection camouflage method, electronic device, and medium based on EOT transformation.

[0005] In a first aspect, embodiments of the present invention provide an active optical projection camouflage method based on EOT transform, the method comprising: Construct an adaptive color transformation EOT algorithm; Configure an adaptive projection compensation strategy based on ambient light feedback; An adaptive projection compensation strategy based on ambient light feedback is used to process the texture pattern using the adaptive color transformation EOT algorithm to obtain the projection pattern. Several sampling points are selected on the surface of the target object, and the position of each sampling point is detected in real time to estimate the pose of the target object; based on the pose of the target object, the projection pattern is projected onto the surface of the target object to achieve camouflage.

[0006] In a second aspect, embodiments of the present invention provide an electronic device, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs that can be executed by the at least one processor, and the one or more computer programs are executed by the at least one processor to enable the at least one processor to perform the above-described active optical projection camouflage method of EOT transformation.

[0007] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described active optical projection camouflage method for EOT transformation.

[0008] Fourthly, embodiments of the present invention provide a computer program product, including a computer program / instruction, which, when executed by a processor, implements the above-described active optical projection camouflage method for EOT transformation.

[0009] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides an active optical projection camouflage method based on EOT transform. It utilizes EOT to model the geometric and photometric transformations of the target during imaging. By dynamically adjusting the color mapping function, it adaptively corrects color shifts during projection, ensuring the color distribution of the projected pattern is as close as possible to the expected value, thereby improving the camouflage effect. Furthermore, during projection, a detection camera facing the target area is added to detect ambient light in real time. An adaptive projection compensation strategy based on ambient light feedback is designed to dynamically adjust the brightness and contrast of the projected pattern. Combined with target tracking, the projection area is updated in real time, effectively addressing the effects of deformation, changes in illumination, and target motion. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of the active optical projection camouflage method based on EOT transformation of the present invention; Figure 2 This is a schematic diagram illustrating the image test results of mud turtles according to the present invention; Figure 3 This is a schematic diagram of the test results of river bird images according to the present invention; Figure 4 This is a schematic diagram of the test results of the adaptive projection compensation strategy of the present invention; Figure 5 This is a schematic diagram of the test results for recognizing moving target objects according to the present invention; Figure 6 This is a schematic diagram of an electronic device provided by the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] It should be noted that, unless otherwise specified, the features in the following embodiments and implementation methods can be combined with each other.

[0014] like Figure 1 As shown, this embodiment of the invention provides an active optical projection camouflage method based on EOT transformation, the method comprising the following steps: Step S1: Construct the adaptive color transformation EOT algorithm.

[0015] Specifically, under controlled conditions, a set of ambient light intensity E, projection distance D, and optimal color transformation parameters were collected. This allows for the training of a regression model relating the optimal color transformation parameters to the ambient light intensity E and the projection distance D; the expression for the regression model is as follows: ; ; ; ; in, , , , For regression function, These represent the slope and intercept of the hue transformation, respectively. These represent the slope and intercept of the saturation transformation, respectively.

[0016] Obtain real-time ambient light intensity and real-time projection distance Based on the trained regression model, the current optimal color transformation parameters are obtained; based on the current optimal color transformation parameters, the texture pattern T is mapped to the projection pattern P; The EOT algorithm is used to perform a geometric transformation on the projected pattern P to obtain the projected pattern P'.

[0017] Current optimal color transformation parameters The expression is as follows: ; ; ; ; In the formula, These represent the slope and intercept of the current hue transformation, respectively. The current values ​​represent the slope and intercept of the saturation transformation, respectively. Based on the current optimal color transformation parameters, the texture pattern T is mapped to the projection pattern P, as shown in the following expression: ; ; ; In the formula, H and S represent the hue and saturation components of the texture pattern T in the HSV color space, respectively; H represents hue, S represents saturation, and C represents the color transformation function.

[0018] Step S2: Configure the adaptive projection compensation strategy for ambient light feedback.

[0019] Ambient lighting images are acquired at a frequency of f (fps). In this example, f=15~30; and the acquired RGB image is converted to the CIEXYZ color space, and the Y channel is extracted as the luminance image. Calculate the brightness image The mean value is denoted as ambient light intensity. , which serves as a measure of ambient light intensity at time t, and is used for subsequent global brightness adjustment.

[0020] According to ambient light intensity Dynamically adjust ambient lighting image The corrected ambient lighting image is obtained. The expression is as follows:

[0021] In the formula, α and β represent weights. , , and These represent the lower and upper limits of brightness, respectively. This refers to ambient light intensity; when the ambient light is relatively dim ( Increase brightness when the ambient light is bright; when the ambient light is bright ( Reduce brightness when ). and The values ​​are set based on the statistical values ​​of ambient light reflectance (EM) and shape factor (FM).

[0022] Obtain physical brightness distribution According to physical brightness distribution Find the corrected ambient lighting image Areas where clipping may occur This means that local brightness compensation is performed on areas where the brightness exceeds the shape factor or is lower than the ambient light reflectance value. Only in the region The local brightness is adjusted using the following formula:

[0023] In the formula, yes Highlighted areas outside the FM range yes Dark areas below EM. For bright areas, reduce local brightness as much as possible without exceeding FM; for dark areas, reduce brightness without falling below the corresponding JND threshold. Under the premise of moderately increasing brightness, avoid clipping.

[0024] Finally, smooth with a Gaussian filter. The adjustment results for the region show that the smoothing radius is inversely proportional to the high-frequency component R_t, ensuring the continuity of the adjustment in the high-frequency region and reducing visibility.

[0025] For the corrected ambient lighting image Radiation compensation is performed using the following expression:

[0026] In the formula, Indicates the ambient light reflectance. Represents the shape factor; It should be noted that, The ambient light reflectance value represents the intensity of ambient light received by the camera at the (x, y) pixel location without the projector being turned on, reflecting the impact of ambient lighting conditions on the projection effect. The shape factor combines the reflectivity of the projection surface and the value of the projector-surface shape factor at the (x, y) pixel position, reflecting the attenuation of the projected light intensity due to the projection distance and projection angle, and characterizing the spatial non-uniformity of the projection system.

[0027] Ambient lighting image after radiation compensation Time-domain smoothing is performed using the following expression:

[0028] In the formula, Indicates the weight.

[0029] Step S3: Based on the adaptive projection compensation strategy of ambient light feedback, the texture pattern is processed by the adaptive color transformation EOT algorithm to obtain the projection pattern.

[0030] Step S4: Select several sampling points on the surface of the target object, detect the position of each sampling point in real time, and thus estimate the pose of the target object; based on the pose of the target object, project the projection pattern onto the surface of the target object to achieve camouflage.

[0031] The process of selecting several sampling points on the surface of the target object, detecting the position of each sampling point in real time, and thus estimating the pose of the target object includes: Several sampling points are selected on the surface of the target object; assuming that n sampling points are uniformly selected on the surface of the target object, with coordinates as follows: , .

[0032] Establish a local coordinate system for the sampling points and obtain the relationship between the local coordinates and the global coordinates; select a point O on the object's surface as the origin of the local coordinate system, and establish the local coordinate system based on the object's shape characteristics. Sampling points The local coordinates are The transformation relationship between local and global coordinates is as follows:

[0033] In the formula, R is The rotation matrix, T is Translation vectors describe rigid body transformations from the local coordinate system to the global coordinate system. 0 represents vectors with all elements equal to zero. vector.

[0034] The position of each sampling point is detected in real time, and the coordinates of each sampling point in the global coordinate system are obtained. Based on the position of each sampling point and the relationship between local and global coordinates, the pose of the target object is estimated. This is achieved using the coordinates of the sampling points in the local coordinate system. and the global coordinates obtained by measurement A least squares problem can be constructed to estimate the transformation matrix. :

[0035] Solve using singular value decomposition. The obtained R and T represent the current position and orientation of the object.

[0036] Furthermore, the object's position and orientation are updated in real time. Due to environmental disturbances and the object's own motion, its position and orientation change over time. By continuously tracking sampling points and estimating the transformation matrix, the object's position and orientation can be updated in real time. The update frequency depends on the sensor's sampling frequency and the efficiency of the estimation algorithm.

[0037] Resampling may be necessary. If the object has undergone significant deformation, the original sampling points may no longer be applicable. In this case, it is necessary to re-select sampling points on the object's surface and repeat the above steps. The frequency of resampling depends on the specific application.

[0038] Example 1 This example demonstrates an active light projection camouflage method based on EOT transform, applied to camouflage images of mud turtles and river birds in the ImageNet dataset. First, the original images are recognized using a ResNet18 image recognition model based on a convolutional neural network. Then, adversarial projection patterns are generated using the proposed adaptive color transformation EOT algorithm.

[0039] The generated adversarial projection pattern is fused with the original image and then input into the ResNet18 model for recognition. Experimental results show that the adversarial projection generated using the Adaptive Color Transformation (EOT) algorithm can effectively deceive the image recognition model, causing it to produce incorrect classification results. In the attack on mud turtle images, the model misclassifies the adversarial examples as chitons, boletes, etc., achieving a 100% attack success rate. Figure 2 As shown. In attacks on river bird images, the model misclassifies adversarial examples as brain corals, partridges, etc., achieving an attack success rate of 80%, such as... Figure 3 As shown.

[0040] Example 2 Using the Fruit30 dataset, a projection attack on fruit images is implemented in the physical world through an adaptive projection compensation strategy. First, a pre-trained Fruit30 fruit recognition model is used to identify the original fruit images. Then, an ambient light detection camera is installed near the projector to acquire real-time images of the ambient light in the projection area, and the projection compensation coefficient is dynamically adjusted based on the ambient light intensity and the brightness of the projected pattern.

[0041] In physical-world projection attack tests, adversarial projection patterns were projected onto the surfaces of different types of fruit using an EPSON projector, and tests were conducted under varying lighting conditions. Experimental results show that applying an adaptive projection compensation strategy can effectively improve the contrast and stability of the projected patterns, maintaining good visual quality even in complex lighting environments. In indoor environments, the projection attack success rate reached 95%, causing the Fruit30 fruit recognition model to produce incorrect classification results. In outdoor environments, the projection attack success rate reached 90%, such as... Figure 4 As shown, the feasibility and robustness of this method in practical applications are demonstrated.

[0042] Example 3 A pre-trained ResNet18 model was used to identify moving objects. In the experiments, different categories of objects, such as books, water glasses, and toys, were selected, and their movement was simulated by manually moving or rotating them. Simultaneously, an Intel RealSense depth camera was used to track sampling points on the object's surface in real time, and the projection pattern was dynamically adjusted based on changes in the sampling point's position.

[0043] Experimental results show that the real-time update strategy can effectively adapt to the movement of the target object, ensuring that the projected pattern always maintains consistency with the position and shape of the target object. Even when the target object undergoes translation, rotation, or partial occlusion, the success rate of projection camouflage remains above 85%. Figure 5 As shown, compared with traditional static projection camouflage methods, this method significantly improves the robustness and adaptability of camouflage, providing a feasible solution for achieving projection camouflage in dynamic scenes.

[0044] Accordingly, this application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; and, when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the active optical projection camouflage method based on EOT transformation as described above. Figure 6 The diagram shown is a hardware structure diagram of any device with data processing capabilities for the active optical projection camouflage method based on EOT transformation provided in this embodiment of the invention, except... Figure 6 In addition to the processor, memory, and network interface shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0045] Accordingly, this application also provides a computer-readable storage medium storing computer instructions thereon, which, when executed by a processor, implement the active light projection camouflage method based on EOT transformation as described above. The computer-readable storage medium can be an internal storage unit of any data processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data processing device, and can also be used to temporarily store data that has been output or will be output.

[0046] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0047] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. An active optical projection camouflage method based on EOT transform, characterized in that, The method includes: Construct an adaptive color transformation EOT algorithm; Configure an adaptive projection compensation strategy based on ambient light feedback; An adaptive projection compensation strategy based on ambient light feedback is used to process the texture pattern using the adaptive color transformation EOT algorithm to obtain the projection pattern. Several sampling points are selected on the surface of the target object, and the position of each sampling point is detected in real time to estimate the pose of the target object; based on the pose of the target object, the projection pattern is projected onto the surface of the target object to achieve camouflage.

2. The active optical projection camouflage method based on EOT transformation according to claim 1, characterized in that, The process of constructing the adaptive color transformation EOT algorithm includes: Under controlled conditions, a set of ambient light intensity E, projection distance D, and optimal color transformation parameters were collected. This allows for the training of a regression model between the optimal color transformation parameters and the ambient light intensity E and projection distance D. Obtain real-time ambient light intensity and real-time projection distance Based on the trained regression model, the current optimal color transformation parameters are obtained; based on the current optimal color transformation parameters, the texture pattern T is mapped to the projection pattern P; The EOT algorithm is used to perform a geometric transformation on the projected pattern P to obtain the projected pattern P'.

3. The active optical projection camouflage method based on EOT transformation according to claim 2, characterized in that, The expression for the regression model is as follows: ； ； ； ； in, , , , For regression function, These represent the slope and intercept of the hue transformation, respectively. These represent the slope and intercept of the saturation transformation, respectively.

4. The active optical projection camouflage method based on EOT transformation according to claim 3, characterized in that, Obtain real-time ambient light intensity and real-time projection distance Based on the trained regression model, the current optimal color transformation parameters are obtained; The process of mapping texture pattern T to projection pattern P based on the current optimal color transformation parameters includes: Current optimal color transformation parameters The expression is as follows: ； ； ； ； In the formula, These represent the slope and intercept of the current hue transformation, respectively. The current values ​​represent the slope and intercept of the saturation transformation, respectively. Based on the current optimal color transformation parameters, the texture pattern T is mapped to the projection pattern P, as shown in the following expression: ； ； ； In the formula, H and S represent the hue and saturation components of the texture pattern T in the HSV color space, respectively; H represents hue, S represents saturation, and C represents the color transformation function.

5. The active optical projection camouflage method based on EOT transformation according to claim 1, characterized in that, The process of configuring an adaptive projection compensation strategy based on ambient light feedback includes: Acquire ambient lighting images ; According to ambient light intensity Dynamically adjust ambient lighting image The corrected ambient lighting image is obtained. The expression is as follows: ; In the formula, α and β represent weights. and These represent the lower and upper limits of brightness, respectively. Ambient light intensity; Obtain physical brightness distribution According to physical brightness distribution Find the corrected ambient lighting image Local brightness compensation is applied to areas where the brightness exceeds the shape factor or is lower than the ambient light reflectance. For the corrected ambient lighting image Radiation compensation is performed using the following expression: ; In the formula, Indicates the ambient light reflectance. Represents the shape factor; Ambient lighting image after radiation compensation Time-domain smoothing is performed using the following expression: ; In the formula, Indicates the weight.

6. The active optical projection camouflage method based on EOT transformation according to claim 1, characterized in that, The process of selecting several sampling points on the surface of the target object, detecting the position of each sampling point in real time, and thus estimating the pose of the target object includes: Select several sampling points on the surface of the target object; Establish a local coordinate system for the sampling points and obtain the relationship between the local coordinates and the global coordinates; The position of each sampling point is detected in real time, and the coordinates of each sampling point in the global coordinate system are obtained. Based on the position of each sampling point and the relationship between local and global coordinates, the pose of the target object is estimated.

7. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs that can be executed by the at least one processor, the one or more of the computer programs being executed by the at least one processor to enable the at least one processor to perform the active optical projection camouflage method of EOT transformation as described in any one of claims 1-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the active optical projection camouflage method of EOT transformation as described in any one of claims 1-6.

9. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the active optical projection camouflage method of EOT transformation as described in any one of claims 1-6.

Images