Method and device for identifying occlusion space of three-dimensional field of view based on equidistant gradient shadow

Abstract

The present application relates to the field of video monitoring, in particular to a three-dimensional field-of-view occlusion space identification method and device based on equidistant gradient shadow. The present application can accurately identify the occlusion profile and accurately calculate the shadow of the occlusion profile by efficiently judging and analyzing the field-of-view occlusion area, so as to quickly obtain the space model of the field-of-view occlusion obstacle area, and truly present the actual visual range and effective coverage space of the camera. Moreover, the present application can realize the simulation of video monitoring by software, ensure the authenticity and accuracy of the simulation results, and improve the work efficiency of the designers. The simulation results can be used to scientifically guide the optimization of the design and construction of video monitoring, and ensure to meet the requirements of full video coverage without dead angle in key public places.

Description

Technical Field

[0001] This invention relates to the field of video surveillance, and in particular to a method and device for three-dimensional field-of-view occlusion space recognition based on equally spaced gradient shadows. Background Technology

[0002] Video surveillance, as a crucial technological means in the security industry, has become a common user requirement due to increasingly complex monitoring environments and rising demands for public security. However, limited by current technology, to meet this requirement, engineering design and implementation often resort to a simplistic approach of deploying cameras at multiple points to supplement each other's field of view. This has led to a series of pressing problems, including unreasonable deployment schemes, high construction and implementation costs, excessive duplication of monitoring, and cumbersome data processing and system maintenance.

[0003] Due to the lack of practical and feasible methods for in-depth analysis of the field of view monitoring range, traditional two-dimensional planar methods or the rapidly developing three-dimensional BIM and other engineering design methods struggle to accurately calculate the occlusion effects of other objects within the field of view, thus failing to obtain the effective monitoring range and field of view coverage of a single camera. Therefore, a three-dimensional field of view occlusion spatial identification method is needed to quickly and accurately obtain the effective monitoring range and field of view coverage of a camera. Summary of the Invention

[0004] The purpose of this invention is to overcome the problem in the prior art that it is difficult to accurately calculate the occlusion effect of other objects within the field of view, thus making it impossible to obtain the effective monitoring range and field of view coverage of a single camera, and to provide a three-dimensional field of view occlusion space recognition method and device based on equally spaced gradient shadows.

[0005] To achieve the above-mentioned objectives, the present invention provides the following technical solution:

[0006] A method for spatial recognition of occlusion in a 3D field of view based on equally spaced gradient shadows includes the following steps:

[0007] S1: Construct the camera's field of view model;

[0008] S2: Detect occlusions within the camera's field of view along the inner normal of the field of view, and output the set of regions where occlusions exist.

[0009] S3: Obtain the trapezoidal offset of each occluded area along the outer normal of the camera's field of view;

[0010] S4: Calculate the corresponding trapezoidal spacing based on the trapezoidal offset of each occluded area, and create gradient planes step by step to cut the occluded areas into equally spaced trapezoidal cuts.

[0011] S5: Based on the equally spaced trapezoidal cut results of each occlusion area, calculate the shadow contour of the relevant occlusion object on the corresponding gradient plane;

[0012] S6: Sequentially merge and create shadow contours on adjacent gradient planes to form the camera's field of view occlusion obstacle areas caused by each occluder and its shadow.

[0013] As a preferred embodiment of the present invention, the camera field-of-view model in S1 is a three-dimensional parametric model with an arc-shaped bottom and a cone shape, and the maximum value of its field-of-view radius L is the camera's farthest effective monitoring distance L. max .

[0014] As a preferred embodiment of the present invention, step S2 includes the following steps:

[0015] S21: Calculate the step spacing:

[0016] L step =L max / N;

[0017] Among them, L step L is the step spacing, N is the preset step size, and L is the step size. max This refers to the camera's maximum effective monitoring distance.

[0018] S22: Let the field of view radius L = n × L step Collision detection is performed within the current camera's field of view.

[0019] If a collision occurs, it means that there is an obstruction in the current field of view of the camera;

[0020] If there are intermittent consecutive collisions, it indicates that there are multiple occlusion areas within the current camera's field of view;

[0021] Otherwise, there are no obstructions in the current camera field of view model;

[0022] Where n is the step value, n∈[1, 2, 3, ..., N];

[0023] S23: If there are continuous collisions or intermittent continuous collisions, the step value n of the first collision is marked as the starting step value of the current occluded area, and the step value n of the last collision is marked as the ending step value of the current occluded area.

[0024] S24: Set the step value n = n + 1, and proceed to S22; until the starting and ending step values ​​of all occluded areas within the camera's field of view are obtained.

[0025] As a preferred embodiment of the present invention, step S3 includes the following steps:

[0026] S31: Select the occlusion area A along the normal direction within the camera's field of view.m m is the index of the occluded area, with an initial value of 1;

[0027] S32: Obtain the starting point O of the camera's field of view model s Distance S me ×L step The center point O of the arc surface mc Spatial coordinates;

[0028] Among them, S me For the occluded area A m The termination step value;

[0029] S33: Along the direction perpendicular to the camera's field of view normal, at the center point O of the arc surface. mc Create an offset test plane P′ at the location me Detect offset test plane P′ me Is there an obstruction with a cross-sectional profile?

[0030] S34: If there is no obstruction on the cross-sectional profile, let S me =S me -1, K m =K m +1, enter S32;

[0031] If an obstruction exists, output the current K. m For the occluded area A m The trapezoidal surface offset is entered into S35;

[0032] Among them, K m This is the offset of the currently occluded area, with an initial value of 0;

[0033] S35: Let m = m + 1, then proceed to S31 until the trapezoidal offset of all occluded areas is obtained.

[0034] As a preferred embodiment of the present invention, step S4 includes the following steps:

[0035] S41: Select the occlusion area A along the normal direction within the camera's field of view. m Where m is initially 1;

[0036] S42: Calculate the occlusion area A m The spacing between the trapezoidal surfaces, including the offset of the trapezoidal surfaces, is D. mi And along the normal direction within the camera's field of view, obtain the coordinates with the starting point O of the camera's field of view model. s The distance is the spacing between the trapezoidal surfaces, D. mi The center point O of the arc surface mi Spatial coordinates;

[0037] Among them, D mi =(S mi-K m )×L step S mi ∈[S ms S me ], its initial value is S ms ;

[0038] S43: Along the direction perpendicular to the normal of the camera's field of view, at the center point O of the arc surface. mi Create a gradient plane P perpendicular to the camera's field of view normal. mi ;

[0039] S44: Let S mi =S mi +1, enter S42, until S mi =S me Complete the shading of area A m Equally spaced trapezoidal cuts;

[0040] S45: Let m = m + 1, then proceed to S41 until the equidistant trapezoidal cuts of all occluded areas are completed;

[0041] S46: Obtaining and O s Distance L max The center point O of the arc surface ec Spatial coordinates, in O ec Create a gradient plane P perpendicular to the field of view normal. e .

[0042] As a preferred embodiment of the present invention, step S5 includes the following steps:

[0043] S51: Traverse the equally spaced cut trapezoidal surfaces of M occlusion regions and connect them with the gradient plane P. e The gradient plane set P of the camera field of view model is stored together, and the step value corresponding to each gradient plane in the gradient plane set P is stored in the step value set S.

[0044] Wherein, the gradient plane set P = {P1, ..., P2} i , ..., P e} and the set of step values ​​S = {S1, ..., S2} i S e The elements within the} correspond sequentially;

[0045] S52: Along the inner normal direction of the camera's field of view, sequentially obtain each gradient plane P. i The outline of the cross section of the obstruction on the upper part OL i ;

[0046] S53: Get the cross-sectional contour of the occluded object (OL) iThe information of the model lines is used to calculate the control points of each model line and the starting point O of the field of view model. s Spacing D ij ;

[0047] The information of the model line includes the line type and control points OP. ij Coordinates, radius, and angle; OP ij For gradient plane P i The j-th control point;

[0048] S54: Calculate OP ij In the gradient plane P i+1 The shaded point OP' ij Starting point O of the camera's field of view model s Spacing D′ ij ;

[0049] Among them, D′ ij =D ij ×S i+1 / S i ;

[0050] S55: In O s Starting from OP, ij Placed on the ray vector of O s Spacing is D′ ij The shaded point OP' ij ;

[0051] S56: Repeat S54 and S55 until the outline of the occluded object section OL is completed. i Place the shadow points of all control points in the model lines;

[0052] S57: Based on the cross-sectional outline of the obstruction OL i The model lines are of the line type, and all their corresponding shadow points are connected in sequence to form the cross-sectional outline OL of the occluding object. i In the gradient plane P i+1 The shadow line OL′ on i ;

[0053] S58: Obtain the gradient plane P i+1 Upper section profile OL i+1 closed-loop region PA i+1 and shaded line OL′ i The closed-loop region PA′ i Calculate PA i+1 and PA′ i The union region is output, and the outline of the union region is used as the occlusion region on the gradient plane P. i+1 The cross-sectional profile OL on i+1 ;

[0054] S59: Repeat S53 to S58 until the shadow contours of all occluders on the corresponding gradient plane are obtained.

[0055] As a preferred embodiment of the present invention, in S57, when the model line is of the arc type, the drawing radius R′=R×S i+1 / S i , where R is the radius of the model line.

[0056] As a preferred embodiment of the present invention, step S6 includes the following steps:

[0057] S61: Select the cross-sectional contour lines on adjacent gradient planes in sequence, and continuously create hollow body models using the fusion method;

[0058] S62: Connects all hollow body models and outputs the occlusion areas of the camera's field of view caused by each occluder and its shadow.

[0059] As a preferred embodiment of the present invention, S6 further includes S63;

[0060] S63: Connect the solid model of the camera's field of view and the hollow model of the obstruction area, and output the actual field of view coverage model of the camera.

[0061] A three-dimensional field-of-view occlusion space recognition device based on equally spaced gradient shadows includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to perform any of the methods described above.

[0062] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0063] This invention efficiently analyzes and judges the occlusion area of ​​the field of view, accurately identifies the outline of the obstruction, and precisely calculates the shadow of the obstruction outline, thereby quickly obtaining a spatial model of the occlusion obstacle area and realistically presenting the actual visible range and effective coverage space of the camera. Furthermore, this invention can simulate video surveillance coverage through software, ensuring the authenticity and accuracy of the simulation results and improving the work efficiency of designers. The output simulation results can be used to scientifically guide the optimization of video surveillance design and construction, ensuring that the requirement for complete video coverage without blind spots in key public places is met. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating a three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows as described in Embodiment 1 of the present invention.

[0065] Figure 2This is a three-dimensional schematic diagram of the camera field of view parameterized BIM component with arc bottom and cone solid shape in a three-dimensional field of view occlusion space recognition method based on equidistant gradient shadows as described in Embodiment 4 of the present invention.

[0066] Figure 3 This is a schematic diagram of the elevation of the camera field of view parameterized BIM component with the bottom of the arc surface and the shape of the cone solid in a three-dimensional field of view occlusion space recognition method based on equidistant gradient shadows as described in Embodiment 4 of the present invention.

[0067] Figure 4 This is a schematic diagram of the camera's field of view and the occlusion scene in a three-dimensional field of view occlusion space recognition method based on equally spaced gradient shadows as described in Embodiment 4 of the present invention;

[0068] Figure 5 This is a schematic diagram of the detection of occlusion objects by the equal-interval step distance of the inner normal in a three-dimensional field-of-view occlusion space recognition method based on equal-interval gradient shadows as described in Embodiment 4 of the present invention.

[0069] Figure 6 This is a schematic diagram showing that the outline of the occluded object cannot be captured at the cross-section of the center point of the arc surface in a three-dimensional field-of-view occlusion space recognition method based on equidistant gradient shadows as described in Embodiment 4 of the present invention.

[0070] Figure 7 This is a schematic diagram illustrating the process of obtaining the cutting offset by equidistant translation of the outer normal in a three-dimensional field-of-view occlusion space recognition method based on equidistant gradient shadows as described in Embodiment 4 of the present invention.

[0071] Figure 8 This is a schematic diagram of an equal-spacing offset cutting occlusion object in a three-dimensional field-of-view occlusion space recognition method based on equal-spacing gradient shadows as described in Embodiment 4 of the present invention;

[0072] Figure 9 This is a schematic diagram of picking and drawing the cross-sectional contour of the occlusion object on the gradient plane in a three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows as described in Embodiment 4 of the present invention.

[0073] Figure 10 This is a schematic diagram of the cross-sectional contour shadow of the gradient plane obtained in a three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows as described in Embodiment 4 of the present invention.

[0074] Figure 11 This is a schematic diagram of the inner normal superposition and trimming of the upper gradient plane shadow in a three-dimensional field of view occlusion space recognition method based on equidistant gradient shadows as described in Embodiment 4 of the present invention.

[0075] Figure 12This is a schematic diagram of the contour shadow of the occlusion region on the Pe gradient plane in a three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows as described in Embodiment 4 of the present invention.

[0076] Figure 13 This is a schematic diagram of the hollow body of the built-in field-of-view occlusion region in a three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows as described in Embodiment 4 of the present invention.

[0077] Figure 14 This is a schematic diagram of the field-of-view occlusion space recognition method based on equidistant gradient shadows described in Embodiment 4 of the present invention, in which hollow bodies are connected to form a field-of-view occlusion obstacle area.

[0078] Figure 15 This is a schematic diagram of the structure of a three-dimensional field-of-view occlusion space recognition device based on equidistant gradient shadows, which utilizes the three-dimensional field-of-view occlusion space recognition method based on equidistant gradient shadows described in Embodiment 1 or 2 of the present invention. Detailed Implementation

[0079] The present invention will be further described in detail below with reference to experimental examples and specific embodiments. However, this should not be construed as limiting the scope of the above-mentioned subject matter of the present invention to the following embodiments; all technologies implemented based on the content of the present invention fall within the scope of the present invention.

[0080] Example 1

[0081] like Figure 1 As shown, a three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows includes the following steps:

[0082] S1: Construct the camera's field of view model.

[0083] S2: Detect occlusions within the camera's field of view along the inner normal of the field of view, and output a set of regions where occlusion exists.

[0084] S3: Obtain the trapezoidal offset of each occluded area along the outer normal of the camera's field of view.

[0085] S4: Calculate the corresponding trapezoidal spacing based on the trapezoidal offset of each occluded area, and create gradient planes step by step to cut the occluded areas into equally spaced trapezoidal surfaces.

[0086] S5: Based on the equally spaced trapezoidal cut results of each occlusion area, calculate the shadow contour of the relevant occlusion object on the corresponding gradient plane.

[0087] S6: Sequentially merge and create shadow contours on adjacent gradient planes to form the camera's field of view occlusion obstacle areas caused by each occluder and its shadow.

[0088] Example 2

[0089] This embodiment is a specific implementation of the three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows described in Embodiment 1, including the following steps:

[0090] S1: Construct the camera's field of view model.

[0091] The camera's field-of-view model is a three-dimensional parametric model with an arc-shaped bottom and a cone shape. The maximum value of its field-of-view radius L is the camera's farthest effective monitoring distance L. max .

[0092] S2: Detects obstructions within the camera's field of view along the inner normal and outputs a set of regions where obstructions exist. Specifically, it detects obstructions within the effective viewing distance along the inner normal of the camera's field of view and outputs a set of obstructed regions including the start and end step values.

[0093] S21: Calculate the step spacing:

[0094] L step =L max / N;

[0095] Among them, L step Where N is the step spacing and N is the preset step size.

[0096] S22: Let the field of view radius L = n × L step Collision detection is performed within the current camera's field of view.

[0097] If a collision occurs, it means that there is an obstruction in the current field of view of the camera;

[0098] If there are intermittent consecutive collisions, it indicates that there are multiple occlusion areas within the current camera's field of view;

[0099] Otherwise, there are no obstructions within the current camera's field of view;

[0100] Where n is the step value, n∈[1,2,3,...,N].

[0101] S23: If there are continuous collisions or intermittent continuous collisions, mark the step value n of the first collision as the starting step value of the current occluded area, and mark the step value n of the last collision as the ending step value of the current occluded area.

[0102] S24: Set the step value n = n + 1, then proceed to S22 until the starting and ending step values ​​of all occluded areas within the camera's field of view are obtained.

[0103] S3: Obtain the trapezoidal offset of each occluded area along the outer normal of the camera's field of view.

[0104] S31: Select the occlusion area A along the normal direction within the camera's field of view. m m is the index of the occluded area, with an initial value of 1.

[0105] S32: Obtain the starting point O of the camera's field of view model s Distance S me ×L step The center point O of the arc surface mc Spatial coordinates.

[0106] Among them, S me For the occluded area A m The termination step value.

[0107] S33: Along the direction perpendicular to the camera's field of view normal, at the center point O of the arc surface. mc Create an offset test plane P′ at the location me Detect offset test plane P′ me Does the cross-sectional profile of the obstruction exist?

[0108] S34: If there is no obstruction on the cross-sectional profile, let S me =S me -1, K m =K m +1, enter S32;

[0109] If an obstruction exists, output the current K. m For the occluded area A m The trapezoidal surface offset is entered into S35;

[0110] Among them, K m This is the offset of the currently occluded area, with an initial value of 0.

[0111] S35: Let m = m + 1, then proceed to S31 until the trapezoidal offset of all occluded areas is obtained.

[0112] S4: Calculate the corresponding trapezoidal surface spacing based on the trapezoidal surface offset of each occluded area, and create gradient planes step by step to cut each occluded area with equal spacing. That is, based on the trapezoidal surface offset, the starting step value, and the ending step value, create gradient planes step by step, and cut each occluded area perpendicularly at equal intervals along the camera's field of view normal.

[0113] S41: Select the occlusion area A along the normal direction within the camera's field of view. m ; where the initial value of m is 1.

[0114] S42: Calculate the occlusion area A mThe spacing between the trapezoidal surfaces, including the offset of the trapezoidal surfaces, is D. mi And along the normal direction within the camera's field of view, obtain the coordinates with the starting point O of the camera's field of view model. s The distance is the spacing between the trapezoidal surfaces, D. mi The center point O of the arc surface mi Spatial coordinates;

[0115] Among them, D mi =(S mi -K m )×L step S mi ∈[S ms S me ], its initial value is S ms .

[0116] S43: Along the direction perpendicular to the normal of the camera's field of view, at the center point O of the arc surface. mi Create a gradient plane P perpendicular to the camera's field of view normal. mi .

[0117] S44: Let S mi =S mi +1, enter S42, until S mi =S me Complete the shading of area A m Equally spaced trapezoidal cuts.

[0118] S45: Let m = m + 1, then proceed to S41 until all equidistant trapezoidal cuts of the occluded areas are completed.

[0119] S46: Obtaining and O s Distance L max The center point O of the arc surface ec Spatial coordinates, in O ec Create a gradient plane P perpendicular to the field of view normal. e .

[0120] S5: Based on the equally spaced trapezoidal cut results of each occluding region, calculate the shadow contour of the relevant occluding object on the corresponding gradient plane. That is, obtain the cross-sectional contour of each gradient plane, calculate the shadow contour of the cross-sectional contour on the next gradient plane, and form the cross-section and shadow contour of the next gradient plane.

[0121] S51: Traverse the equally spaced cut trapezoidal surfaces of M occlusion regions and connect them with the gradient plane P. e The gradient plane set P of the camera field of view model is stored together, and the step value corresponding to each gradient plane in the gradient plane set P is stored in the step value set S.

[0122] Wherein, the gradient plane set P = {P1, ..., P2} i, ..., P e} and the set of step values ​​S = {S1, ..., S2} i S e The elements within the} are sequentially matched.

[0123] S52: Along the inner normal direction of the camera's field of view, sequentially obtain each gradient plane P. i The outline of the cross section of the obstruction on the upper part OL i .

[0124] S53: Get the cross-sectional contour of the occluded object (OL) i The information of the model lines is used to calculate the control points of each model line and the starting point O of the field of view model. s Spacing D ij .

[0125] The information of the model lines includes line type (such as arc properties, straight line properties, etc.) and control points OP. ij (e.g., starting point, ending point, center of circle, etc.) coordinates, radius, and angles (e.g., starting angle, ending angle, etc.); OP ij For gradient plane P i The j-th control point.

[0126] S54: Calculate OP ij In the gradient plane P i+1 The shaded point OP' ij Starting point O of the camera's field of view model s Spacing D′ ij .

[0127] Among them, D′ ij =D ij ×S i+1 / S i (That is, using the trigonometric ratio formula, the spacing D′ is obtained by the ratio of the step values ​​of the trapezoidal surfaces.) ij ).

[0128] S55: In O s Starting from OP, ij Placed on the ray vector of O s Spacing is D′ ij The shaded point OP' ij .

[0129] S56: Repeat S54 and S55 until the outline of the occluded object section OL is completed. i Place the shadow points of all control points in the model lines.

[0130] S57: Based on the cross-sectional outline of the obstruction OL i The model lines are of the line type, and all their corresponding shadow points are connected in sequence to form the cross-sectional outline OL of the occluding object.i In the gradient plane P i+1 The shadow line OL′ on i .

[0131] When the model line is of circular arc type, the drawing radius R′=R×S i+1 / S i R is the radius of the model line.

[0132] S58: Obtain the gradient plane P i+1 Upper section profile OL i+1 closed-loop region PA i+1 and shaded line OL′ i The closed-loop region PA′ i Calculate PA i+1 and PA′ i The union region is output, and the outline of the union region is used as the occlusion region on the gradient plane P. i+1 The cross-sectional profile OL on i+1 .

[0133] S59: Repeat S53 to S58 until the shadow contours of all occluders on the corresponding gradient plane are obtained.

[0134] S6: Sequentially fuse and create shadow contours on adjacent gradient planes to form occlusion obstacle areas in the camera's field of view caused by various occluders and their shadows. That is, sequentially fuse the cross sections and shadow contours of adjacent gradient planes to create spatial models of various occluders and their shadow areas, completing the identification and acquisition of the occlusion space in the camera's field of view.

[0135] S61: Select the cross-sectional contour lines on adjacent gradient planes in sequence, and continuously create hollow body models using the fusion method.

[0136] S62: Connects all hollow body models and outputs the occlusion areas of the camera's field of view caused by each occluder and its shadow.

[0137] Example 3

[0138] The difference between this embodiment and the previous embodiment is that S6 further includes S63;

[0139] S63: Subtract the occlusion area of ​​the camera's field of view from the camera's field of view model to output the effective field of view model of the camera.

[0140] Example 4

[0141] This embodiment is a practical application of the three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows described in Embodiment 3, including the following steps:

[0142] S1. In the Revit family library environment, set the furthest effective monitoring distance L. max As a parameter of the arc radius, a camera field of view parameterized BIM family type is constructed for the bottom of the arc and the shape of the cone solid, such as... Figure 2 and Figure 3 As shown.

[0143] S201. Deploy a camera field of view family instance in the Revit project environment and set the maximum effective monitoring distance L. max The field of view is 100m; within this camera's field of view, draw sphere 1, cuboid 1, and sphere 2 made of non-transparent material, which will serve as occlusion 1, occlusion 2, and occlusion 3 respectively. Figure 4 As shown.

[0144] S202, with the step size N set to 200, calculate the step spacing L. step It is 0.5m.

[0145] S203. The current step value n increases from 1 to 200, and the field of view model L is set sequentially. max The value is n × 0.5m; L is adjusted each time. max Then, it is determined whether there is a collision detection between the field of view model and the occluded object.

[0146] S204. When the current step value n is 92, the first collision detection occurs. The collision overlap is located outside the field of view of the occluder 2, marked as occlusion region A1. Record the initial step value S of A1. 1s The value is 92, and the collision overlap intersection U is obtained. n .

[0147] S205, Continue stepping n, update the collision, overlap, and intersection set U. n and with U n-1 A comparison is performed. If a change occurs, the collision continues; otherwise, it indicates that there are no newly added occluded objects in the current occluded area, and the termination step value S of the occluded area is recorded. e For n-1, such as S 1e The value is 104.

[0148] S206. Referring to S204 and S205, detect the occlusion area A2 and record S. 2s S 2e They are 150 and 176 respectively.

[0149] S301. Due to the randomness of the orientation of obstructions within the field of view, the vertical section of the field of view normal at the center point of the bottom arc surface of the current field of view model has... Figure 6 The phenomenon shown indicates that the outline of the occluded object cannot be captured. Therefore, it is necessary to obtain the trapezoidal offset of the occluded area, such as... Figure 7 As shown.

[0150] S302. Along the normal direction within the field of view, pick the point O at the starting point of the field of view model. s The spacing is 52m (S) 1e ×L step The center point O of the arc surface 1c .

[0151] S303, in O 1c Create test plane P′ at [location]. 1e There is no obstruction to the cross-sectional profile; it is necessary to offset L along the out-of-field normal direction. step Continue creating the test plane P′ at 0.5m. 1e-1 At this point, a cross-sectional profile exists. The offset K1 of the occluded region A1 is recorded as 1, and plane P′ is deleted. 1e-1 、P′ 1e .

[0152] S304, referring to S302 and S303, create test planes sequentially until the cross-sectional profile appears. Obtain the offset K2 of the occluded region A2 as 4, and delete plane P′. 2e-4 、P′ 2e .

[0153] S401. Along the normal direction within the field of view, pick the point O at the starting point of the field of view model. s The spacing is 45.5m ((S) 1s -K1)×L step The center point O of the arc surface 11 .

[0154] S402, along the direction perpendicular to the field of view normal, at O 11 Create a gradient plane P perpendicular to the field of view normal. 101 .

[0155] S403, refer to S401 and S402, such as Figure 8 As shown, gradient planes P for occlusion regions A1 and A2 are created sequentially. 102 ~P 113 P 201 ~P 227 All occluded areas within the field of view are uniformly divided at equal intervals along the field of view direction. Specifically, A1 and A2 are divided into 13 and 27 parts, respectively.

[0156] S404, the pickup field of view radius is L max Center point O of the bottom arc surface of the time field model ec Create the bottom gradient plane P according to S402. e .

[0157] S501. Sequentially select and draw the gradient plane P. 101 ~P 113 P201 ~P 227 The outline of the cross section of the obstruction on the upper part OL 101 ~OL 113 OL 201 ~OL 227 ,like Figure 9 As shown.

[0158] S502, according to OL 101 Model line type, pick OL in sequence 101 All model line control points constitute OP. 101 The point set is used to calculate the starting point O of the field of view model. s To OP 101 The distance D between points in the middle j and the direction of the ray vector.

[0159] S503, given that the field-of-view model step detection spacing and occlusion region cutting spacing are both L... step Furthermore, the overall offset spacing of the cut trapezoidal surface in the obscured area is L. step The mechanism of integer multiples allows us to use the trigonometric ratio formula to calculate OP. 101 Each point at P 102 The shadow point on the field of view model starting point O s Spacing D′ j =D j ×92 / 91.

[0160] S504, along OP 101 Ray vector direction at each point and the spacing D′ between shaded points j In P 102 Place the starting point O of the field of view model on top s The light source is blocked by the OP point. 101 The resulting set of shaded points OP′ 101 ,like Figure 10 As shown.

[0161] S505, according to OP 101 The line types of the model lines are connected sequentially to draw OP'. 101 The points in the middle form OP 101 In P 102 The shaded line OL′1.

[0162] S506. Determine the relationship between OL′1 and OL2. Since it is a full inclusion relationship, first delete OL2 and then set OL′1 as P. 102 The cross-sectional profile line on the surface.

[0163] S507. Referring to S502 to S506, draw the shading lines sequentially and update P. 102 ~P 113 P 201~P 227 and P e The cross-sectional profile, such as Figure 11 , Figure 12 As shown.

[0164] S601, sequentially pick up OL on the adjacent gradient planes i and OL i+1 A hollow volume model is built into the field of view model using a fusion method, such as... Figure 13 As shown.

[0165] S602, connect the regions that obstruct the field of view, and then cut them to obtain a camera field of view model that can realistically represent the effective visible range, such as... Figure 14 As shown.

[0166] Example 5

[0167] like Figure 15 As shown, a three-dimensional field-of-view occlusion space recognition device based on equally spaced gradient shadows includes at least one processor, a memory communicatively connected to the at least one processor, and at least one input / output interface communicatively connected to the at least one processor. The memory stores instructions executable by the at least one processor, which, when executed, enables the at least one processor to perform the three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows described in the foregoing embodiments. The input / output interface may include a display, keyboard, mouse, and USB interface for inputting and outputting data.

[0168] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0169] When the integrated units of this invention are implemented as software functional units and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, 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 cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0170] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows, characterized in that, Includes the following steps: S1: Construct the camera's field of view model; S2: Detect occlusions within the camera's field of view along the inner normal of the field of view, and output the set of regions where occlusions exist. S3: Obtain the trapezoidal offset of each occluded area along the outer normal of the camera's field of view; S4: Calculate the corresponding trapezoidal spacing based on the trapezoidal offset of each occluded area, and create gradient planes step by step to cut the occluded areas into equally spaced trapezoidal cuts. S5: Based on the equally spaced trapezoidal cut results of each occlusion area, calculate the shadow contour of the relevant occlusion object on the corresponding gradient plane; S6: Sequentially merge and create shadow contours on adjacent gradient planes to form the camera's field of view occlusion obstacle areas caused by each occluder and its shadow; S3 includes the following steps: S31: Select the occlusion area A along the normal direction within the camera's field of view. m m is the index of the occluded area, with an initial value of 1; S32: Obtain the starting point O of the camera's field of view model s Distance S me ×L step The center point O of the arc surface mc Spatial coordinates; Among them, S me For the shading area A m The termination step value; S33: Along the direction perpendicular to the camera's field of view normal, at the center point O of the arc surface. mc Create an offset test plane P′ at the location. me Detect offset test plane P′ me Is there an obstruction with a cross-sectional profile? S34: If there is no obstruction on the cross-sectional profile, let S me =S me -1, K m =K m +1, enter S32; If an obstruction exists, output the current K. m For the shading area A m The trapezoidal surface offset is entered into S35; Among them, K m This is the offset of the currently occluded area, with an initial value of 0; S35: Let m = m + 1, then proceed to S31 until the trapezoidal offset of all occluded areas is obtained.

2. The method for three-dimensional field-of-view occlusion space recognition based on equally spaced gradient shadows according to claim 1, characterized in that, In S1, the camera's field of view model is a three-dimensional parametric model with an arc-shaped bottom and a cone shape. The maximum value of its field of view radius L is the camera's farthest effective monitoring distance L. max .

3. The method for three-dimensional field-of-view occlusion space recognition based on equally spaced gradient shadows according to claim 1, characterized in that, S2 includes the following steps: S21: Calculate the step spacing: L step =L max / N; Among them, L step L is the step spacing, N is the preset step size, and L is the step size. max This refers to the camera's maximum effective monitoring distance. S22: Let the field of view radius L = n × L step Collision detection is performed within the current camera's field of view. If a collision occurs, it means that there is an obstruction in the current field of view of the camera; If there are intermittent consecutive collisions, it indicates that there are multiple occlusion areas within the current camera's field of view; Otherwise, there are no obstructions within the current camera's field of view; Where n is the step value, n∈[1, 2, 3, ..., N]; S23: If there are continuous collisions or intermittent continuous collisions, the step value n of the first collision is marked as the starting step value of the current occluded area, and the step value n of the last collision is marked as the ending step value of the current occluded area. S24: Set the step value n=n+1, and proceed to S22; until the starting and ending step values ​​of all occluded areas within the camera's field of view are obtained.

4. The method for three-dimensional field-of-view occlusion space recognition based on equally spaced gradient shadows according to claim 1, characterized in that, S4 includes the following steps: S41: Select the occlusion area A along the normal direction within the camera's field of view. m Where m is initially 1; S42: Calculate the occlusion area A m The spacing between the trapezoidal surfaces, including the offset of the trapezoidal surfaces, is D. mi And along the normal direction within the camera's field of view, obtain the coordinates with the starting point O of the camera's field of view model. s The distance is the spacing between the trapezoidal surfaces, D. mi The center point O of the arc surface mi Spatial coordinates; Among them, D mi =(S mi -K m )×L step S mi ∈[S ms S me ], its initial value is S ms ; S43: Along the direction perpendicular to the normal of the camera's field of view, at the center point O of the arc surface. mi Create a gradient plane P perpendicular to the camera's field of view normal. mi ; S44: Let S mi =S mi +1, enter S42, until S mi =S me Complete the shading of area A m Equally spaced trapezoidal cuts; S45: Let m = m + 1, then proceed to S41 until the equidistant trapezoidal cuts of all occluded areas are completed; S46: Obtaining and O s Distance L max The center point O of the arc surface ec Spatial coordinates, in O ec Create a gradient plane P perpendicular to the field of view normal. e .

5. The three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows according to claim 1, characterized in that, S5 includes the following steps: S51: Traverse the equally spaced cut trapezoidal surfaces of M occlusion regions and connect them with the gradient plane P. e The gradient plane set P of the camera field of view model is stored together, and the step value corresponding to each gradient plane in the gradient plane set P is stored in the step value set S. Wherein, the gradient plane set P = {P1, ..., P2} i , ..., P e } and the set of step values ​​S = {S1, ..., S2} i S e The elements within the} correspond sequentially; S52: Along the inner normal direction of the camera's field of view, sequentially obtain each gradient plane P. i The outline of the cross section of the obstruction on the upper part OL i ; S53: Get the cross-sectional contour of the occluded object (OL) i The information of the model lines is used to calculate the control points of each model line and the starting point O of the field of view model. s Spacing D ij ; The information of the model line includes the line type and control points OP. ij Coordinates, radius, and angle; OP ij For gradient plane P i The j-th control point; S54: Calculate OP ij In the gradient plane P i+1 The shaded point OP' ij Starting point O of the camera's field of view model s Spacing D′ ij ; Among them, D′ ij =D ij ×S i+1 / S i ; S55: In O s Starting from OP, ij Placed on the ray vector at O s Spacing is D′ ij OP's shaded point ij ; S56: Repeat S54 and S55 until the outline of the occluded object section OL is completed. i Place the shadow points of all control points in the model lines; S57: Based on the cross-sectional outline of the obstruction OL i The model lines are of the line type, and all their corresponding shadow points are connected in sequence to form the cross-sectional outline OL of the occluding object. i In the gradient plane P i+1 The shadow line OL′ on i ; S58: Obtain the gradient plane P i+1 Upper section profile OL i+1 closed-loop region PA i+1 and shaded line OL′ i The closed-loop region PA′ i Calculate PA i+1 and PA′ i The union region is output, and the outline of the union region is used as the occlusion region on the gradient plane P. i+1 The cross-sectional profile OL i+1 ; S59: Repeat S53 to S58 until the shadow outlines of all occluders on the corresponding gradient plane are obtained.

6. The three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows according to claim 5, characterized in that, In S57, when the model line type is an arc, the drawing radius R′=R×S i+1 / S i , where R is the radius of the model line.

7. The method for three-dimensional field-of-view occlusion space recognition based on equally spaced gradient shadows according to claim 1, characterized in that, S6 includes the following steps: S61: Select the cross-sectional contour lines on adjacent gradient planes in sequence, and continuously create hollow body models using the fusion method; S62: Connects all hollow body models and outputs the occlusion areas of the camera's field of view caused by each occluder and its shadow.

8. A three-dimensional field-of-view occlusion space recognition method based on equally spaced gradient shadows according to claim 7, characterized in that, S6 also includes S63; S63: Connect the solid model of the camera's field of view and the hollow model of the obstruction area, and output the actual field of view coverage model of the camera.

9. A three-dimensional field-of-view occlusion space recognition device based on equally spaced gradient shadows, characterized in that, It includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 8.

Images