CN103093423A - Method of improving spatial feature similarity of screen surface and background space - Google Patents
Method of improving spatial feature similarity of screen surface and background space Download PDFInfo
- Publication number
- CN103093423A CN103093423A CN2012105027213A CN201210502721A CN103093423A CN 103093423 A CN103093423 A CN 103093423A CN 2012105027213 A CN2012105027213 A CN 2012105027213A CN 201210502721 A CN201210502721 A CN 201210502721A CN 103093423 A CN103093423 A CN 103093423A
- Authority
- CN
- China
- Prior art keywords
- camouflage screen
- similarity
- screen surface
- camouflage
- background
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method of improving a spatial feature similarity of a screen surface and a background space space. The method includes the steps of training generalized regression neural network by obtaining target background space sample data, calculating altitude data of a camouflage screen surface feature point by using trained generalized regression neural network with coordinates of the camouflage screen surface feature point as an input vector and obtaining the spatial feature comprehensive similarity of a camouflage screen surface and the background space by analyzing spatial feature parameters of the camouflage screen surface, when the spatial feature comprehensive similarity meets requirements, the camouflage screen surface is produced according to the altitude data of the camouflage screen surface feature point. The altitude data of the camouflage screen surface feature point is produced due to the fact that a targeted background space sample point is collected, the camouflage screen surface which is produced after the spatial feature comprehensive similarity of the camouflage screen surface and the targeted background space is analyzed and evaluated is good in effect of being mixed with the background space and suitable for producing large-area screen surface of various background spaces.
Description
Technical field
The present invention relates to a kind of camouflage screen and look unfamiliar into method, particularly a kind of method that improves shield face and spatial context characteristic similarity.
Background technology
Recognition technology based on image characteristic analysis is present topmost target identification method.This method distinguishes target image according to the statistical discrepancy of target image and background image from background image, and and then reach identification target purpose.Even think that target has and the on all four Color Statistical characteristic of background, horizontal shield and background characteristics distance are still very remarkable.Therefore be necessary to consider the space characteristics of background and target when implementing camouflage.
Summary of the invention
The object of the present invention is to provide a kind of raising shield face of estimating based on generalized regression nerve networks and comprehensive similarity and the method for spatial context characteristic similarity, be intended to improve the space characteristics similarity of camouflage screen face and target background.
The present invention is achieved in that a kind of method that improves shield face and spatial context characteristic similarity, comprises the following steps:
1) obtain coordinate and the corresponding altitude figures of background sample point and the described background sample point of target;
2) with the coordinate of the described background sample point input vector as generalized regression nerve networks, with the altitude figures of the correspondence target vector as generalized regression nerve networks output, described generalized regression nerve networks is trained;
3) generalized regression nerve networks of utilizing training to complete as input vector, calculates the altitude figures of the unique point of camouflage screen face with the coordinate of the unique point of camouflage screen face;
4) analyze the space characteristic parameter of camouflage screen face, draw the space characteristics similarity of camouflage screen face and background, and adopt force definition to draw the space characteristics comprehensive similarity of camouflage screen face and background according to described space characteristics similarity;
5) when described space characteristics comprehensive similarity meets the requirements, generate the camouflage screen face according to the altitude figures of the unique point of described camouflage screen face, otherwise return to step 2), repeat above step.
Described space characteristic parameter comprises average height, variance, roughness, waviness and shade function.
The contour map that utilizes laser range finder, transit or recorded obtains coordinate and the corresponding altitude figures of the sample point of described background and described sample point in conjunction with the resolution of reconnaissance equipment.
The present invention utilizes the generalized regression nerve networks manual intervention few, and the higher characteristics of stability by the background sample point of the target that gathers, generate the altitude figures of camouflage screen face unique point; And the space characteristics comprehensive similarity of the background of evaluating objects and camouflage screen face, differentiate the syncretizing effect of the background of camouflage screen face and target according to described comprehensive similarity, regeneration camouflage screen face when syncretizing effect meets the requirements, thereby the camouflage screen face that generates, good with the background syncretizing effect, the large tracts of land shield that is applicable to all kinds of backgrounds is looked unfamiliar into.
Description of drawings
Fig. 1 is the process flow diagram of the method for the raising shield face that provides of the embodiment of the present invention and spatial context characteristic similarity;
Fig. 2 a-illustraton of model of bell membership function under different values when Fig. 2 c is the computed altitude similarity.
Embodiment
In order to make purpose of the present invention, technical scheme and advantage clearer, below in conjunction with accompanying drawing, the present invention is further elaborated.
Referring to Fig. 1, the figure shows the flow process of the method for raising shield face that the embodiment of the present invention provides and spatial context characteristic similarity.For convenience of explanation, only show the part relevant with the embodiment of the present invention.
A kind of method that improves shield face and spatial context characteristic similarity comprises the following steps:
1) obtain the background sample data of target;
Namely obtain coordinate and the corresponding altitude figures of background sample point and the described background sample point of target;
In the embodiment of the present invention, the contour map that can utilize laser range finder, transit or record is in conjunction with coordinate and the corresponding altitude figures of resolution to obtain described background sample point and described background sample point of reconnaissance equipment;
2）
That is, with the coordinate of the described background sample point input vector as generalized regression nerve networks, with the altitude figures of the correspondence target vector as generalized regression nerve networks output, described General Neural Network is trained;
3) generalized regression nerve networks of utilizing training to complete, the altitude figures of the unique point of calculating camouflage screen face;
The generalized regression nerve networks that mainly refers to utilize training to complete as input vector, calculates the altitude figures of the unique point of camouflage screen face with the coordinate of the unique point of camouflage screen face;
4) analyze the space characteristic parameter of camouflage screen face and background, draw the space characteristics similarity of camouflage screen face and background, and adopt force definition to draw the space characteristics comprehensive similarity of camouflage screen face and background according to described space characteristics similarity;
In the embodiment of the present invention, described space characteristic parameter comprises average height, variance, roughness, waviness and shade function.
5) described space characteristics comprehensive similarity is reviewed, when described space characteristics comprehensive similarity meets the requirements, is generated the camouflage screen face according to the altitude figures of the unique point of camouflage screen face, otherwise return to step 2), repeat above step.
Below, in step 4), calculate the space characteristic parameter that obtains the camouflage screen face, and the calculating of space characteristics similarity and space characteristics comprehensive similarity is elaborated as follows:
The space characteristics of camouflage screen face is explained by following five parameters: average height, reflection be the index of camouflage screen face height, adopt following formula (1) to calculate; Variance, reflection be the index of camouflage screen face height change, adopt following formula (2) to calculate; Roughness, the index of reflection camouflage screen face fluctuations adopts following formula (3) to calculate; Waviness reflects the poor of camouflage screen face maximum elevation and minimum elevation, adopts following formula (4) to calculate; The shade function, the ratio of reflection camouflage screen face exposure area and the total area of investigation, adopts following formula (5) calculating:
（2）
（5）
In formula: be x, the coordinate points data after the gridding of y feeling the pulse with the finger-tip mark,
Refer to the height at coordinate points (x, y) some place, m, n are respectively the sum of putting on latter two direction of gridding,
Be the surface area of shield face unit,
Be the projected area of shield face unit on surface level,
Be maximum elevation,
Be minimum elevation,
Shield face bin slope probability density function,
With
Respectively incident angle and reflection angle,
With
Be respectively during from the metering of x and y direction, meet the bin slope of conditioned reflex,
With
It is respectively the root mean square statistics slope of surface undulation when investigating from x and y direction.
The computing method of height similarity: because most of average heights all have continuity, adopt the average discrepancy in elevation to judge the similarity of shield face and background average height, but because its average discrepancy in elevation of dissimilar background is not quite similar, can utilize a kind of " bell " membership function to come the computed altitude similarity:
Wherein,
Represent the average discrepancy in elevation
Similarity; E represents the truth of a matter of natural logarithm;
, k and r be the parameter of control function curved shape, Fig. 2 gets bell membership function illustraton of model under different parameters; Wherein, in Fig. 2 a, the value of parameter is respectively
, k=2.72, r=1, in Fig. 2 b,
, k=2.72, r=1; In Fig. 2 c,
, k=2.72, r=2.
The computing method of shade similarity: comprise the evaluation of shape relation two aspects of distance between curve, curve, the distance between shield face surface shaded function curve can utilize Euclidean distance formula (7) to calculate distance between each curve
d, its formula is as follows:
Camouflage screen face surface shaded function shape relation can utilize formula of correlation coefficient (8) to calculate, and its formula is as follows:
Distance between comprehensive above-mentioned curve
dAnd camouflage screen face surface shaded function shape relation, give a weighted value for each factor, finally draw the shade similarity.
The computing formula of other index similarity:
Wherein,
,
,
Represent respectively Variance Similarity, roughness similarity and waviness similarity.
Backgroud Df,
Backgroud R,
Backgroud RfRepresent respectively variance, roughness and the waviness of the background of target;
Surface Df,
Surface R,
Surface Rf Represent respectively variance, roughness and the waviness of camouflage screen face.
Comprehensive similarity sim
synCan adopt force definition to give weighted value of five indexs and calculate, see formula (12):
the present invention utilizes the generalized regression nerve networks manual intervention few, the characteristics that stability is higher, by gathering the background sample point of target, generate the data of shield face unique point, then evaluating objects background and camouflage screen space of planes characteristic quantity, draw the similarity of each characteristic quantity, draw at last the degrees of fusion that a comprehensive similarity judges camouflage screen face and target background, the fusion of comprehensive similarity higher shield face and background is better, thereby, the camouflage screen face that uses this technology to generate, good with the background syncretizing effect, the large tracts of land shield that is applicable to all kinds of backgrounds is looked unfamiliar into, have significant military affairs and economic benefit.
The above is only the preferred embodiment of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the principle of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (3)
1. a method that improves shield face and spatial context characteristic similarity, is characterized in that, comprises the following steps:
1) obtain coordinate and the corresponding altitude figures of background sample point and the described background sample point of target;
2) with the coordinate of the described background sample point input vector as generalized regression nerve networks, with the altitude figures of the correspondence target vector as generalized regression nerve networks output, described generalized regression nerve networks is trained;
3) generalized regression nerve networks of utilizing training to complete as input vector, calculates the altitude figures of the unique point of camouflage screen face with the coordinate of the unique point of camouflage screen face;
4) analyze the space characteristic parameter of camouflage screen face, draw the space characteristics similarity of camouflage screen face and background, and adopt force definition to draw the space characteristics comprehensive similarity of camouflage screen face and background according to described space characteristics similarity;
5) when described space characteristics comprehensive similarity meets the requirements, generate the camouflage screen face according to the altitude figures of the unique point of described camouflage screen face, otherwise return to step 2), repeat above step.
2. the method for raising shield face according to claim 1 and spatial context characteristic similarity, is characterized in that, described space characteristic parameter comprises average height, variance, roughness, waviness and shade function.
3. the method for raising shield face according to claim 1 and 2 and spatial context characteristic similarity, it is characterized in that, the contour map that utilizes laser range finder, transit or recorded obtains coordinate and the corresponding altitude figures of the background sample point of described target and described background sample point in conjunction with the resolution of reconnaissance equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012105027213A CN103093423A (en) | 2012-11-30 | 2012-11-30 | Method of improving spatial feature similarity of screen surface and background space |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012105027213A CN103093423A (en) | 2012-11-30 | 2012-11-30 | Method of improving spatial feature similarity of screen surface and background space |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103093423A true CN103093423A (en) | 2013-05-08 |
Family
ID=48205959
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012105027213A Pending CN103093423A (en) | 2012-11-30 | 2012-11-30 | Method of improving spatial feature similarity of screen surface and background space |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103093423A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276753A (en) * | 2019-06-20 | 2019-09-24 | 兰州理工大学 | Objective self-adapting hidden method based on the mapping of feature space statistical information |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102024145A (en) * | 2010-12-01 | 2011-04-20 | 五邑大学 | Layered recognition method and system for disguised face |
-
2012
- 2012-11-30 CN CN2012105027213A patent/CN103093423A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102024145A (en) * | 2010-12-01 | 2011-04-20 | 五邑大学 | Layered recognition method and system for disguised face |
Non-Patent Citations (2)
| Title |
|---|
| EDWARD J. RZEMPOLUCK: "Neural network classification of EEG during camouflaged object identification", 《INTERNATIONAL JOURNAL OF MEDICAL INFORMATICS》, vol. 44, no. 3, 31 May 1997 (1997-05-31), pages 169 - 175 * |
| 瞿珏等: "一种伪装阵地构造方法研究", 《电光与控制》, vol. 17, no. 11, 30 November 2010 (2010-11-30), pages 30 - 34 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276753A (en) * | 2019-06-20 | 2019-09-24 | 兰州理工大学 | Objective self-adapting hidden method based on the mapping of feature space statistical information |
| CN110276753B (en) * | 2019-06-20 | 2021-07-23 | 兰州理工大学 | Target self-adaptive hiding method based on feature space statistical information mapping |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111681197B (en) | Remote sensing image unsupervised change detection method based on Siamese network structure | |
| CN104536009B (en) | Above ground structure identification that a kind of laser infrared is compound and air navigation aid | |
| CN105247461B (en) | Pitching and yaw are determined for touch screen interaction | |
| CN110569796A (en) | Method for dynamically detecting lane line and fitting lane boundary | |
| CN104751185B (en) | SAR image change detection based on average drifting genetic cluster | |
| CN106651886A (en) | Cloud image segmentation method based on superpixel clustering optimization CNN | |
| Jin et al. | A point-based fully convolutional neural network for airborne LiDAR ground point filtering in forested environments | |
| CN104200471B (en) | SAR image change detection based on adaptive weight image co-registration | |
| CN103177451B (en) | Based on the self-adapting window of image border and the Stereo Matching Algorithm of weight | |
| CN102426700B (en) | Level set SAR image segmentation method based on local and global area information | |
| CN103279957A (en) | Method for extracting remote sensing image interesting area based on multi-scale feature fusion | |
| CN105608417A (en) | Traffic signal lamp detection method and device | |
| CN103593679A (en) | Visual human-hand tracking method based on online machine learning | |
| CN103810725A (en) | Video stabilizing method based on global optimization | |
| CN107146238A (en) | The preferred motion target tracking method of feature based block | |
| CN106228569A (en) | A kind of fish speed of moving body detection method being applicable to water quality monitoring | |
| CN110083915A (en) | A kind of high frequency region Radar Target Scatter central forward automation modeling method | |
| Chen et al. | Multi-level interpolation-based filter for airborne LiDAR point clouds in forested areas | |
| CN104484679B (en) | Non- standard rifle shooting warhead mark image automatic identifying method | |
| CN114120150B (en) | Road target detection method based on unmanned aerial vehicle imaging technology | |
| CN113901931B (en) | Behavior recognition method of infrared and visible light video based on knowledge distillation model | |
| CN107341449A (en) | A kind of GMS Calculation of precipitation method based on cloud mass changing features | |
| CN111812670B (en) | Single photon laser radar space transformation noise judgment and filtering method and device | |
| CN109815784A (en) | A kind of intelligent method for classifying based on thermal infrared imager, system and storage medium | |
| CN117765006A (en) | Multi-level dense crown segmentation method based on unmanned aerial vehicle image and laser point cloud |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130508 |