Exeley Inc.

Jianguo, Wang

Ying, Zhu

Application of Incremental Updating Association Mining Algorithm in Geological Disasters System

Xiaoping, Li

Songze, Lei

Yanhong, Wang

Feng, Xiao

Tian, Penghui

Fast Aerial UAV Detection Based on Image Segmentation and HOG-FLD Feature Fusion

Jun, Yu

Zhonghua, Li

Zhiyi, Hu

Jun, Dai

2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
126 
 
The Research and Implementation of 3D Scene Simulation of Camouflage
Yu Jun 
School of computer science and engineering 
Xi'an Technological University 
Xi'an, 710021, ShanXi, China 
e-mail: yujun@xatu.edu.cn 
Li Zhonghua 
School of computer science and engineering 
Xi'an Technological University 
Xi'an, 710021, ShanXi, China 
e-mail: 761173763@qq.com 
Hu Zhiyi 
Engineering Design Institute 
Army Academy of PLA 
Beijing, 100000, China 
e-mail: huzhiyi016v7@163.com 
Dai Jun 
School of computer science and engineering 
Xi'an Technological University 
Xi'an, 710021, ShanXi, China 
e-mail: daijun24@qq.com 
 
Abstract—Aiming at the problems in the camouflage design of 
military engineering, such as strong subjectivity and 
unpredictable camouflage effect, etc. We propose a 3D 
camouflage scene simulation method based on MFC and Vega 
Prime technologies. On the basis of real DEM data in an area 
in American, we use 3D visualization technique to create 
terrain model and ground object model. The camouflage 
pattern is applied to the surface of the model by texture 
mapping technology, and real-time rendering is used. The 
engine Vega Prime API renders the model and creates a 
camouflage simulation scenario. The result shows that, the 
virtual scene of camouflage can be generated by using visual 
simulation technology, this helps to test the effect of the 
camouflage pattern. 
Keywords-Camouflage; Terrain Modeling; Visual 
Simulation; Digital Elevation Model 
I. INTRODUCTION 
Camouflage, as the most basic military camouflage 
technology, is one of the commonly used method to combat 
investigation and attack in modern warfare[1]. The 
commonly used method of Engineering camouflage is to 
paint camouflage patterns on the surface of the target with 
dope, however, the effect of this kind of camouflage is 
influenced by environmental. This usually requires personnel 
to participate in the evaluation of camouflage effects, so it 
may consume a lot of resources. Many scholars began to use 
the Vega prime-based(VP) visual simulation technology to 
build 3D model of military target or reproduce a realistic 
environment to achieve realistic simulation results. For 
example, Zhang[2] combines the development features of 
MFC application program, gives a 3D visual development 
method based on multi-process technology, and applies it to 
the development of a 3D scene of a portable air defense 
missile simulator software. Yao[3] using the virtual 
battlefield large terrain visualization features and focuses on 
the research of large-scale real terrain visualization 
techniques.Hu[4] made a further improvement on Zhang's 
foundation[2], he gives a program framework that 
seamlessly integrates VP under VC++'s MFC library 
programming environment, optimizing the structure of the 
system. However, none of the above methods can perform 
real-time simulation of real 3D space and do not have 
interactive capabilities, making these methods unsuitable for 
simulation of engineering camouflage. 
In response to these problems, this paper is based on the 
3D visual simulation method[2-6] and real DEM data, 
combined with the actual engineering camouflage 
environment, establish the Vega Prime visual simulation 
framework based on MFC, design a real-time roaming 
system simulation of a camouflage scene. 
II. THE OVERALL DESIGN OF THE SIMULATION SYSTEM 
The camouflage 3D scene simulation system includes 
two parts: simulation model and simulation scene. The 
simulation model refers to all the models in the target area, 
and the simulation scene refers to the virtual camouflage 
simulation scene generated in order to achieve the effect of 
the engineering camouflage simulation.  
For engineering camouflage simulation system design 
and development can be divided into the following four steps 
to complete, and the overall system framework is shown in 
Figure 1. 
 
Terra Vista
Creator
Camouflage 
generate system
terrain model
Culture Data
Surface layout
Texture material
DEM data
ortho-photo data
ground object 
model
Camouflage 
pattern Vega
Custom 
code
API
Real-time 
operation
Real-time 
application
Simulation scene of 
Engineering 
camouflage
 
Figure 1. Framework of Simulation system 
A. Build terrain model 
Acquire high-precision DEM data and satellite 
ortho-photo data of a certain region and establish a terrain 
model of a simulation scene in Terra Vista to construct the 
altitude and fluctuation of the surface. 
B. Build ground object model 
According to the plane layout, texture material and other 
data of the ground surface around the simulation scene, the 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
127 
 
ground model was constructed by MultiGen Creator, thereby 
constructing the scene of military equipment, weapons and 
equipment, and storage camps. 
C. Camouflage scene model 
According to the background and target characteristics 
around the simulation scene, Yu’s[7] camouflage design 
method was used to obtain the digital camouflage pattern, 
the camouflage pattern was mapped to the surface of the 3D 
model, and the disguised model was arranged in the whole 
scene. 
D. Rendering and implementation 
The OpenFlight API is used to programmatically export 
the layout information of the model, then uses Vega Prime 
and visual studio to write the real-time scheduling display 
program, and combines the MFC framework to generate a 
real-time roaming simulation scenario. 
III. ESTABLISH AND RENDERING MODEL 
A. The establishment of terrain model 
In the field of military camouflage, terrain refers to the 
natural state of the earth's surface. Whether a target can be 
detected or not depends on the complexity of the terrain and 
the difference between target and background[8]. Terrain 
model is an abstract and simplified expression of real terrain 
attributes. Therefore, it is inevitably involved in the display 
and processing of the 3D geospatial scenes in the battlefield 
environment simulation. The terrain visual simulation has 
become the hot spot of the 3D scene simulation research in 
the battlefield environment. 
Terra Vista is MultiGen's software for modeling complex 
terrain visual simulations and is used to build 3D terrain 
simulation models. Terra Vista manages 3D terrain data in a 
project. It is mainly composed of a terrain data source library, 
terrain parameter configuration, vector parameter 
configuration, CDB format terrain file browsing, and a 
model database. It runs on the Windows platform and 
generates a terrain database in MataFlight or OpenFlight 
format that can be used directly on PC workstations or other 
graphical vision systems. Terra Vista builds a 3D terrain 
simulation model mainly including three steps: data loading, 
setting terrain parameters, and calculating & generating 
terrain.  
1) data loading 
It mainly includes the import of data such as elevation 
data, satellite image data, and vector data. 
First, load digital elevation data. Take a certain 60km2 
area of America as a camouflage scene. The more accurate 
the DEM data is, the more realistic the terrain is, and the 
more detailed local details of the real terrain can be 
expressed. For simulation scenarios, the accuracy of the 
terrain grid is generally 10m to 30m. Therefore, Yu[9] 
proposed an improved bilinear interpolation algorithm to 
interpolate on the basis of the original DEM data, thereby 
improving the accuracy of DEM data.  
30m accuracy of the original elevation map and Yu's 
method to 15m accuracy elevation map shown in Figure 2. 
 
(a)30m accuracy elevation map (b) 15m accuracy elevation map 
Figure 2. 30m and 15m accuracy elevation map 
Second, load satellite image data. Use Bing Maps 
Satellite to obtain satellite imagery of the area to be 
camouflaged. The image is a Mercator projection on a 
SPHERE RADIUS 6378137 benchmark, and the SRTM 
DEM uses a Geocentric projection of a WGS84 benchmark, 
So this requires format conversion. Use the Projection 
Conversion Tool in Global Mapper to convert satellite 
imagery into the WGS84 benchmark Geocentric projection 
format. The ortho-photos obtained are shown in Figure 3. 
 
 
Figure 3. Ortho-photos maps 
Finally, load vector data. The vector data refers to the 
polygons in the terrain database except terrain. It can be 
formed naturally, such as lakes and rivers, or it can be 
constructed artificially, such as buildings and roads. we use 
satellite image to create vector data for this area by Terra 
Vista. 
The main rules are: to create point details for objects with 
limited area, location, orientation, that do not require a 
specific shape; to create linear details for objects that length 
is much larger than the width and height; to create area 
details for objects that occupy a large area and need to be 
defined on the edge of the area. 
2) Terrain parameter settings 
It mainly includes the levels of detail (LOD), the visual 
distance, the size of the grid, and the method of construct 
grid. Terrain LOD technology is a kind of terrain drawing 
method, which reduces the geometric complexity of the 
terrain scene by simplifying the details of the terrain surface 
step by step without affecting the visual effects of the scene, 
thereby improving the efficiency of the rendering algorithm. 
In this paper, the LOD is constructed using a quadtree 
hierarchy, each layer divides the terrain into several grids, 
the next layer quarters each grid in the previous layer, and so 
on. Each layer in the hierarchy Grid is subdivided by a 
number of triangles. The simulation scene LOD is set to 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
128 
 
three levels, the visual distance is set to 5000m, the terrain 
block is set to 7500m×7500 m, and Denauley algorithm is 
used to construct a triangular mesh to generate 
multi-resolution geometric terrain. AS shown in Figure 4. 
 
 
(a) Terrain model with 582 patches (LOD2) 
 
(b) Terrain model with 8496 patches (LOD1) 
 
(c) Terrain model with 18,171 patches (LOD0) 
Figure 4. Terrain model with different resolutions 
3) Calculate and generate terrain 
Use the Gaming Area tool to select the target area to be 
generated, set the format and resolution of the output texture 
image, determine the output model format, taking into 
account the later amendments in the Creator, this article uses 
OpenFlight format. In the process of generating the terrain 
model, the generated terrain is stored in the OpenFlight 
model file in the unit of divided grid. Finally, each model 
node is unified by an external reference node in the file 
named master.flt. Figure 5 is shown as the final geometric 
terrain. 
 
 
Figure 5. Terrain map 
B. Build ground object model 
The ground object model is an important part of the 
battlefield environment, including natural land objects, 
military equipment, blockhouses, storage barracks and other 
important battlefield objects. MultiGen Creator is a visual 
simulation software for creating 3D models. The basic core 
of the software is the modeling module, which provides a 
visual environment for creating and editing database 
files[10]. Different from common modeling tools such as 
3DMAX and MAYA, Creator uses a hierarchical structure to 
describe the simulation scenario. It can conveniently 
organize the model according to the geometric feature and is 
suitable for real-time traversal operations[11]. Steps for 
using Creator modeling are as follows:  
Step1: Determine model data. For natural objects, 
information such as location and area is required; For 
buildings, they need information such as floor plan, section 
drawing, and three views. 
Step2: Determine the hierarchical structure of the model. 
Decompose the model according to the tree structure, such as 
the floor of the building, the inner wall, etc. 
Step3: Using Yu's algorithm[7] to generate camouflage 
pattern, the selected typical background and the generated 
camouflage pattern are shown in the following page Figure 
6. 
 
(a) typical background (b) camouflage pattern 
Figure 6. The generation of camouflage patterns. 
Step4: Using texture mapping technology to Figure 
6.camouflage pattern mapping to the surface of 3D model, 
through the interactive control module rotating object or 
change the point of view to get a different perspective, has 
practical application value of camouflage target 3D view. 
Finally, the 3D model is placed in the simulation scene 
according to the corresponding proportion. The ground 
object model is shown in Figure 7 and Figure 8. 
 
 
Figure 7. Helicopter camouflage model 
 
Figure 8. Tank camouflage model 
C. Rendering and implementation of the model 
Real-time rendering engine is the core of camouflage 
simulation scene. Vega Prime, a real-time 3D visual 
simulation tool developed by MultiGen, is widely used in 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
129 
 
military simulation[12]. VP includes LynX Prime graphical 
user interface configuration tool and VSG advanced 
cross-platform scene rendering API[13]. This article calls VP 
in the environment of MFC library to implement the design 
of virtual real-time simulation system. The specific method 
is as follows: 
Step1: Creates a single thread in the application to run 
the whole VP program, include initialization, define, 
configure, and frame cycles. 
Step2: Adds the member function responsible for 
opening the thread in the view class of the MFC framework. 
Step3: Uses the API function AfxBegin Thread to open 
the VP thread, and continuously refresh the VP visual 
window in the background program. 
IV. SIMULATION RESULTS AND ANALYSIS. 
The developed engineering camouflage scene system 
based VP technology is shown in Figure 9 The system 
interface is mainly composed of 4 parts. Among them, part 1 
is a subpanel which control to create model, including load 
OpenFlight formatted data of terrain and ground object 
model data, generates digital camouflage pattern, and 
implements the camouflage to terrain and object model in 
the simulation scene; Part 2 is a configuration subpanel of 
rendering, including setting the observer, moves mode and 
atmospheric environment in the simulation, and implements 
the control of simulation scene, such as translation, scaling, 
rotation, etc; Part 3 is a simulation control subpanel, and 
control the beginning and end of the roaming of the whole 
simulation scene; Part 4 is a display subpanel of simulation 
scene, show the full face of the scenario. The user can use 
the PC device to roam freely in the virtual scene. The system 
will perform real-time rendering and updating according to 
the position of the movement. The user can view the layout 
of the simulation scene from different perspectives and 
positions by switching from a close-range perspective to a 
distant perspective. 
 
 
Figure 9. Camouflage scene system interface 
As we have seen from Figure 9, the terrain model and 
object model established by this system are consistent with 
the feature of the regional target background. The simulated 
scene after the disguise has high fidelity and realistic feeling, 
which is not easy to be detected, which fully reflects the 
importance of visual simulation in camouflage. 
V. CONCLUSION 
In this paper, we have improved the traditional 
engineering camouflage method through 3D scene 
simulation technology. we build a framework of VP visual 
simulation based on MFC, to achieve the target area terrain 
and object modeling and rendering, establish a 3D, dynamic 
and high fidelity simulation scene of camouflage. The result 
shows that, the camouflage simulation scene designed in this 
paper visually and truly reproduces the effect of camouflage 
operations in the field, enhances the user's sense of 
immediacy, and solved the camouflage effect evaluation in 
current military engineering mainly depends on the field 
work. It provides an effective auxiliary measure for the 
evaluation of engineering camouflage effect and provides a 
good virtual simulation platform for the design of military 
engineering camouflage, at the same time, this helps to 
reduce resource consumption for camouflage evaluation. 
 
REFERENCES 
[1] Hu Jianghua. Camouflage technology[M]. Beijing, National Defence 
Industry Press, 2012: 84-88. 
[2] Zhang L, Han J Y, Zhang J, et al. Research of Program Development 
Technology of Vega Prime Based on MFC Framework[J]. Fire 
Control & Command Control, 2014, 03(39): 159-162. 
[3] Yao F F, Liang Q, Xu R J, et al. Study of three-dimensional virtual 
battlefield large terrain dynamically generated based on vega prime[J]. 
Journal of System Simulation, 2012, 24(9):1900-1904. 
[4] Zi-Nan H U, Jin-Song Y U. Research of Software Integrated 
Technology of Vega Prime Based on MFC Programming 
Framework[J]. Journal of System Simulation, 2009, 
21(14):4291-4290. 
[5] Wu G, Tao J. 3D Modeling of the Relievo Based on the Computer 
Active Vision[C]// International Conference on Mechatronics, 
Computer and Education Informationization. 2016. 
[6] Liu P, Lu Z. Construction and Visualization of 3D Landscape[C]// 
International Conference on Computer Network, Electronic and 
Automation. IEEE Computer Society, 2017:28-32. 
[7] Yu Jun, Shuang Xiao. Design of Imitation Digital Camouflage[J]. 
Journal of Applied Sciences, 2012, 30(4):331-334. 
[8] Dong Zhiming, Guo Qisheng. Modeling and simulation of battlefield 
environment[M]. Beijing, National Defence Industry Press, 2013: 
153-157. 
[9] Yu Jun, Lu Yanxin, Hu Zhiyi, et al. An Improved Bilinear 
Interpolation of Regular Grid DEM[J]. Journal of Geomatics Science 
and Technology, 2015, 32(5):521-524. 
[10] Shao Xiaodong. MultiGen Creator modeling art[M]. Xi`an, Xidian 
university press, 2014: 40-43. 
[11] Wang N, Han Y, Chen G. Comparison Analysis of Modeling of 
Virtual GIS Based on 3DS MAX and MultiGen Creator[J]. Bulletin 
of Surveying & Mapping, 2010, 62(5):14. 
[12] Vega Prime Programmer’s Guide(Version 2.0)[Z]. Dallas: 
MultiGen-Paradigm Inc. 2005: 12-16. 
[13] Wang Xiaoping. Vega Prime real time 3d virtual reality development 
technology[M]. Chengdu, Southwest Jiaotong University press, 2012: 
128-133. 


The Research and Implementation of 3D Scene Simulation of Camouflage

Nian, Wang

Chen, Hui

Ding, Dawei

A New Non-Singular Terminal Sliding Mode Control and Its Application to Chaos Suppression in Interconnected Power System

Yu, Liu

Junhuai, Zhao

Semantic Analysis of the Wa Tautology in Subordinate Clauses with Minimal Independence Degree

Chaoyang, Geng

Fenli, Gao

2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
44 
 
An Improved Algorithm of the Collision Detection Based on OBB 
Geng Chaoyang 
School of Computer Science and Engineering 
Xi’an Technological University 
Xi’an, 710021, China  
e-mail: gengbox@163.com 
Gao Fenli 
School of Computer Science and Engineering 
Xi’an Technological University 
Xi’an, 710021, China  
e-mail: 1151466533@qq.com 
 
Abstract—Collision detection plays a vital role in improving 
the sense of immersion and realism in virtual environment. 
The bounding box is the most basic collision detection 
algorithm, The OBB intersect test tightness and can able to 
significantly reduce the number of bounding volume. and try 
to occupy less storage space. Propose a bounding box of 
direction cylindrical So it can be improved though surrounded 
objects more close, then improve the efficiency of collision 
detection. 
Keywords-Collision Detection; Detection Algorithms; 
Bounding Box of Direction Cylindrical 
I. THE BASIC PRINCIPLE OF COLLISION DETECTION 
Collision detection is inevitable problems about robot, 
animated simulation, virtual reality, The basic task is to 
determine the two or more objects make contact between 
each other whether or through Real-time and accuracy are 
two important requirements collision detection.  
A. Detection requirements 
1) Test whether there is a collision 
2) The position of the collision detection 
3) Detection distance between objects 
4) Prediction of the collision of the next time 
II. COLLISION DETECTION ALGORITHM CLASSIFICATION 
AND COMPARED 
In the field of collision detection algorithm variety, there 
is no unified classification standards. Here are two aspects of 
collision detection algorithm classification: From the point 
of view of a time domain to points; From the point of view 
of the spatial domain to points. 
A. Detection algorithm based on time domain 
The detection algorithm based on the time domain from 
the point of view of time domain to points, collision 
detection algorithm can be divided into discrete collision 
detection algorithm and continuous collision detection 
algorithm of two. 
Because the algorithm time discrete characteristics, this 
kind of algorithm is at least the following two questions:  
1) Pierce the phenomenon exists. 
2) Missed a collision. 
B. Based on space domain collision detection algorithm 
classification 
Based on space domain collision detection algorithm has 
always been the focus of research, According to the different 
structure they are divided into two types: space division 
method and the hierarchical bounding volume method. the 
space division method is used for the entire sequence level 
division technology to realize, and the hierarchical bounding 
volume law is in the scene of each object of constructing the 
rational hierarchical bounding volume to achieve. The object 
bounding box test is a collision detection algorithm is widely 
used in a way. Common bounding volume of the bounding 
sphere, of AABB of OBB bounding box. 
1) Based on the method of spherical bounding box of 
collision detection 
Based on the method of spherical bounding box of 
collision detection in virtual scene, two irregular objects in 
motion will be the collisions between, can use the spherical 
bounding box method to carry on the examination, adopt 
corresponding measures to avoid a collision. As shown in 
figure 1 first to realize the dynamic pick virtual object, and 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
45 
 
then based on the actual need, for each of the virtual objects 
involves creating spherical bounding box. At two objects, 
get the distance between the two and earn two sphere of 
their respective radius. When testing conditions have been 
met, according to the spherical bounding box collision 
detection algorithm, judge whether the two objects has hit. If 
there was a collision, will conduct collision response, 
according to the current characteristics of the virtual object, 
take different response way. 
Virtual objects up
Create spherical bounding 
box
Centre distance for Get sphere radius
Collision detection
 Collision Detection
 
Figure 1. Spherical bounding box 
2) Based on the coordinate transformation of the 
bounding box 
Axial bounding box AABB is defined as the object 
contains parallel to each side of the coordinate 
transformation of the smallest hexahedral, is the earliest 
application of the bounding box. AABB can be expressed as: 

 
Among them ， ， ，  ， ，  respectively 
is the AABB in the X, Y, Z coordinate the projection on the 
minimum and maximum coordinates. Calculation of the 
object graph respectively of RMB each element of the vertex 
set x, y and z coordinates of the minimum and maximum can 
sure, so describe a AABB only six scalar. 
As shown in figure 2, AABB intersection between the 
test is relatively simple, Abounding box in the projection of 
the x-axis minimum value is  ,A maximum of  B 
bounding box in the projection of  the x-axis minimum 
value is , B maximum of , If > , A 
and B do not intersect, if < , A and B do not 
intersect. If and only if two AABB overlap in the three axes 
of the projection interval, they intersect. 
 
 
 
 
 
 
Figure 2. AABB bounding box 
The AABB advantage of simple structure, easy 
intersection test, computationally efficient, the drawback is 
that the object surrounded by enough compact. 
3) Oriented bounding box (OBB) and the improvement 
of it 
Oriented Bounding Box (OBB) and the improvement of 
its direction the Bounding Box namely OBB ,is to be 
detected and object contains each side, it by Gottschalk put 
forward in 1996. OBB and AABB is the biggest difference 
of the direction of arbitrariness, can be the shape of the 
object structure characteristics of the bounding box cuboids 
any direction, so OBB has better tight sex. The bounding 
box is a very effective collision detection accelerated 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
46 
 
method , based on OBB bounding box here proposes a 
direction cylindrical bounding box. 
As shown in figure 3, direction methods of constructing 
the bounding box cylinder is, first choose need to tear open 
outfit parts in the geometric center O = (ox, oy, oz) for 
vertex, tear open outfit direction v

 = (vx, vy, vz) for 
direction ( v

is the unit vector), tear open outfit distance d 
for length, tectonic straight line, the parameter equation for: 









zz
yy
xx
tvoz
tvoy
tvox


Figure 3. Direction cylinder 
For the parts of the vertex P, set the line for projection, 
the center to the distance to O, traverse parts on all point, 
take the minimum and maximum tmin  tmax, parameters 
interval [tmin， tmax]is parts of all the points on the line of 
projection parameters interval 
So the reference (2. 1) may be given to the parameter 
equation is line for: 








zz
yy
xx
tvoz
tvoy
tvox
(tmin ≤ t ≤ d + tmax)         (3)
Have to line said parts in disassembling movement in 
line in the process of the projection of range, can be used as 
a direction surrounded by the axis of the cylinder box.For 
cylindrical bounding box below the radius. For the parts of 
the vertex P, the P to the center O vectors, point P linear to 
the distance to U=
vOP


 traverse the parts on the 
maximum value umax to the radius r of the cylindrical 
bounding box. 
So far the direction surrounded by cylindrical box has 
been built. 
On the current disassembling parts of the test, at first the 
cylinder of it to all other parts of real intersection test, find 
out of the collision of the parts may occur, and then use the 
disassembling of these parts for accurate parts intersection 
test. 
This premise, given the bounding box of the collision of 
the cylindrical concrete algorithm as follows 
a) Using the method of the introduction to the current 
disassembling parts of cylindrical bounding box structure 
direction, set for axis, radius for r, parts center for O, the 
equation with formula (2.2) said; 
b) The whole equipment traverse the other parts, the 
first part I get the ai, (the first time I = 0, after entering the 
steps of the time every time I since the 1); 
c) Establish a empty set list Li; 
d) To traverse parts of all ai vertex, made the first j a 
point Pij (the first enters, j = 0, every time you go to the 
process after when j since the 1); 
e) For some Pij, its linear to the distance to 
uj=
vOPj


, judge whether uj < r, if found, then turn to step 
6, or turn to step 4; 
f) Will point the triangle strips Pij connected to join set 
Li; 
g) To judge the ai traverse parts all vertices are over, if 
it is to continue step 8, or turn to step 4; 
h) Traverse set Li, will equipment running and each of 
them all triangle triangle strips are surrounded by boxes to 
cylindrical projection on the vertical axis, precise fellowship 
test. 
i) To judge to traverse all parts of the equipment is 
over, if it is to continue step 10, or turn to step 2; 
j) The end of the algorithm.  
Different types of testing methods, the characteristics of 
them are not the same. Surrounded by the ball and AABB 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
47 
 
close degree as flexible Oriented Bounding Box OBB, OBB 
was essentially a closest to the object cuboids, only the 
object of a cuboids’ may, according to the first moment 
arbitrary rotation. OBB surrounded by the ball and AABB 
than closer to the object, It is consistent with the direction, 
construction method , but the gender is poorer, the overall 
OBB performance than AABB surrounded the ball, OBB 
bounding box to the basic idea of the method is to use 
simple geometry instead of complicated geometry novelties, 
first to object bounding box for rough detection, when the 
bounding box intersect surrounded by its geometry can be 
intersection, This can eliminate a large number of 
impossible at the intersection of geometry and geometric 
part, which can quickly find at the intersection of geometric 
part. 
III. CONCLUSION 
Direction surrounded by box of inherited the cylindrical 
OBB bounding box the advantages of the intersection of the 
test in its close the gender is good, can be significantly 
reduced the number of bounding volume, when some similar 
to the cylinder's geometry object happened between the 
collision detection, OBB is a better choice, for direction 
surrounded by box of more suitable for cylinder for 
cylindrical objects geometry.  
Can more closely, more convenient application, so as to 
enhance the efficiency of collision detection. after the rotary 
motion, only need to base the rotation of the coordinate 
system also can. So for the rigid body  
 
REFERENCE 
[1] Sankar Jayaram etal. VADE: A Virtual Assembly Design 
Environment[J]. IEEE Computer Graphics and Applications, 1999(6), 
on page 44-50. 
[2] Ritchie J M， Dewar R G， Simmons J E L. The generation and 
practical use of plans for manual assembly using immersive virtual  
reality [J]. Proc Inst Mech Engrs Part B (S0954-4054), 1999, 23(6): 
461-474. 
[3] Deviprasad T, Kesavadas T. Virtual Prototyping of Assembly 
Components Using Process Modeling [J]. Journal of Manufacturing 
Systems (S0278-6125), 2003, 22(1): 16-27. 
[4] Dewar, R.G., Carpenter, I.D, Ritchie, J.M. Simmons, J.E.L， 
Assembly planning in a virtual environment, PICMET '97: Portland 
International Conference on Management and Technology, Portland, 
1997. 
[5] Jia Zhen Hong, hui Yang, In collision detection problem [J], journal 
of software, 1999, 10(5): 545-551. 
[6] Yu Qin Gao, 3 d space collision detection algorithm of research 
[D] ,Huazhong university of science and technology, master's degree 
thesis, 2007. 
[7] FanZhaoHui, China root, GaoShuMing. Based on parallel fast 
collision detection algorithm [J]. System simulation journal, 2000, 
12(5):548-552. 
[8] Li KangShun, Li YuanXiang, Shang MingDuan, etc.Genetic 
programming in the application of statistical modeling [J]. System 
simulation journal, 2005(7). 
 


An Improved Algorithm of the Collision Detection Based on OBB

Fu, Yanfang

Li, Xing

Feng, Xueyao

Ma, Jing

Advanced Dynamic Autonomous Knowledge Learning Method for Distance Learning

Yufeng, Zhao

Xinwei, Li

Design and Implementation of Music Recommendation System Based on Hadoop

Lian, Zhichao

Wang, Zhongsheng

2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
6 
 
Research on Vehicle Detection Method Based on Background Modeling
Zhichao Lian 
School of Computer Science and Engineering 
Xi’an Technological University 
Xi’an 710032, China 
e-mail: 965941167@qq.com 
Zhongsheng Wang 
School of Computer Science and Engineering 
Xi’an Technological University 
Xi’an 710032, China 
e-mail: wangzhongsheng@xatu.edu.cn 
 
 
Abstract—This paper mainly studies the background 
difference method in the field of intelligent traffic, proposes a 
background modeling method base on frame difference, and 
compares it with the statistical average background model and 
Gaussian distribution background modeling method. vehicle 
contour obtained by the morphological method. Finally, 
experiments were carried out on 4 normal road traffic 
surveillance videos, the effective detection rate used in this 
paper reaches 93.75%, which has a certain degree of 
application. The algorithm model need to be further tested in 
more complex weather and different road conditions. 
Keywords-Vehicle Detection; Background Modeling; 
Inter-frame Difference; Morphological Method 
I. INTRODUCTION 
In practical applications, a vehicle detection method 
with fast response, high accuracy, and good adaptability 
becomes a key part of intelligent traffic detection 
management. Dealing with moving object in video is 
affected by many complicated conditions. Commonly used 
methods for detecting three kinds of moving vehicles 
include: difference method, optical flow method, and 
background difference method. 
Currently, the difference method includes the inter-frame 
difference method and the time difference method. The 
inter-frame difference detection method detectionspeed 
faster and the operation algorithm is simple, and can be 
detected under a scene with high real-time requirements. 
The time difference method is suitable for dynamically 
changing scenes and does not apply to completely 
segmenting moving objects. The optical flow method is 
poor in the context of real-time and practical, and it is 
difficult to meet the requirements for teal-time detection of 
moving vehicles. The background difference method has a 
good result on the speed and detection effect when the 
performance of the camera is relatively stable. The 
background difference method focuses on how to set up the 
background and dynamically update the system in real time. 
This article uses the background difference method. 
II. COMMONLY USED BACKGROUND MODELING UPDATE 
MODELS 
In the monitoring applications, the background 
difference method needs to establish a background reference 
frame. Establishing an accurate and robust background 
model is the key to the system, the accuracy of the frame 
directly affects the output. The commonly used background 
models are the statistical average method and the Gaussian 
distribution background model. 
A. Statistical average method background model 
The statistical average method, also called the mean 
method, is essentially a statistical filter idea. In a period of 
time, the collected images are added together, and the 
average value is taken as the reference background model. It 
is to obtain the gray average value ofN frames of images in 
the sequence image as the estimated value of the 
background image to weaken the interference of the moving 
object to the background. The specific calculation formula 
is shown in 
k 1 2 1
1
( ... )k k k k NAvg f f f f
N
       

Avgk is the background model established for the 
acquisition of the Frame K image system; N is the average 
number of frames; fk,fk-1, ...,fk-N+1 are the continuous 
sequences frame. 
The statistical average method is simple and fast, but if 
easily causes noise accumulation and mixing. This method 
is more suitable for a small number of continuous motion 
objects in the scene, and the background is visible in the 
most of the time. While there are a large number of moving 
objects, especially when they are moving slowly, there will 
be a large deviation in the background of this situation. 
B. Gaussian Distribute Background Model 
The Gaussian distribution background model was first 
proposed by N. Friedman et.al and is divided into 
background models with single Gaussian distribution and 
mixed Gaussian distribution. The single Gaussian model 
method regard the change in the gray value of each pixel in 
the background image as a Gaussian random process, and 
establishes a Gaussian model for each pixel in the image, 
which is achieved by continuously updating the Gaussian 
model background image. The mixed Gaussian model uses 
K (basically 3~5) Gaussian models to characterize the 
features of each pixel in the image. After the new image is 
acquired, the mixed Gaussian model is updated. Each pixel 
in current image is mixed with a Gaussian mixture model. 
Matching to determine the pixel belongs to a background 
point or a front sight. 
The section focuses on the hybrid Gaussian background 
modeling method. This modeling method uses the statistical 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
7 
 
information such as the probability density of a large 
number of sample values for a long time to represent the 
background. Use the statistical difference (such as 3σ 
principle) to judge the target pixel. This method can model 
complex dynamic backgrounds with a large amount of 
computation. Suppose that any pixel (x,y) in the background 
obeys a Gaussian model composed of K Gaussian 
distributions, as shown in follow. 
, , , , , ,
1
( ) ( , , )
k
x y j x y x y j x y j
j
P I I   

 

η(I(x, y), μ(x, y), σ(x,y,j)) means that the j-th Gaussian 
probability density, means value isμ(x,y,j), the variance isσ(x,y,j). 
And the ωjis the weight of the j-th Gaussian distribution.  
The pixel values observed at the current moment are 
compared with the current K Gaussian distribution functions 
in descending order of weights to obtain the best match. If 
there is no match, the pixel is a foreground spot.Otherwise it 
is a background point. The Gaussian distribution 
background model has a large amount of calculation, many 
storage parameters, and takes a long time, which is not 
conducive to practical application. 
III. IMPROVED BACKGROUND MODELING METHOD 
This paper proposes an adaptive background update 
model based on the inter-frame difference method. This 
method uses the background of the current frame and the 
background of the previous frame in the video sequence to 
perform weighted averaging to update the background. The 
specific methods are as shown in formulas (3) and (4) as 
shown. 
( , , ) | ( , , ) ( , , ) |Diff x y t I x y t B x y t 

1 ( , , )
( , , )
0
hDiff x y t T
BOM x y t
otherwise

 
 
I(x, y, t) and B(x, y, t) are the current frame containing 
the moving object at t time and the updated current 
background image; Th is the threshold and the maximum 
peak right fade in the difference image histogram is 
used,and the gray level corresponding to 1/10 of the 
maximum peak. 
Using equation (5) to obtain the motion template 
Stencil(x, y, t) at time t from two adjacent spaced images, it 
is used as a target factor to determine which pixels in the 
current frame are used to update the current background. 
Use the formula (6) to obtain the instantaneous background 
and update the background using the weighted average of 
the instantaneous background and the current background, 
as shown in Equation (7). 
( , , ) ( , , ) & ( , , 1)Stencil x y t BOM x y t BOM x y t 

( , , 1) ( , , ) 0
( , , )
( , , ) ( , , ) 1
temp
B x y t Stencil x y t
B x y t
I x y t Stencil x y t
 
 
 
( , , ) α ( , , ) (1 α) ( , , 1)tempB x y t B x y t B x y t    
Among them, in order to update the coefficient, 
theαvalue has a positive correlation with the update speed. 
The larger theαvalue is, the faster the update speed is, and 
the change of the external light can be captured in time so 
that the current background is closer to the external 
condition of the current frame. The smaller the value is, the 
slower the update rate will be, and the current background 
acquired will have some deviation. After the background 
image is extracted, the current motion area is segmented 
using the background difference method. Using the 
threshold parameter in the expression (8), the image is 
binaries and segmented to obtain the foreground binary 
image. 
1 ( , , )
( , , )
0
Diff x y t threshold
D x y t
other

 
 
Threshold in the formula should be properly selected, 
through the experimental results to filter the residual 
background. 
IV. EXPERIMENTAL TESTING 
According to the flow of the vehicle detection algorithm, 
the original video frame→background update and extraction 
algorithm→motion region segmentation→frame image 
conversion to grayscale image→binary 
processing→morphological processing→obtain the 
detection result, and complete detection test. 
A. Comparison of experimental results 
By using the statistical average method, the background 
of the mixture Gaussian model and the method proposed in 
this paper, the background model is obtained by modeling 
the background image of a video of a traffic video 
surveillance video. The video of this video screen is 1080 * 
960, the frequency is 25fps, and the time length is 10 
seconds.The experimental results are shown are shown in 
Fig.1 below.  
 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
8 
 
 
(a)Original image frame 
The resulting 
background 
frame image 
   
(b-1)Statistical average method (c-1)Mixed Gaussian method (d-1)This article's algorithm 
Background 
and frame 
image 
difference 
processing 
   
(b-2)Statistical average method (c-2)Mixed Gaussian method (d-2)This article's algorithm 
Binary 
operation 
   
(b-3)Statistical average method (c-3)Mixed Gaussian method (d-3)This article's algorithm 
Open 
operation 
   
(b-4)Statistical average method (c-4)Mixed Gaussian method (d-4)This article's algorithm 
Close 
operation 
   
(b-5)Statistical average method (c-5)Mixed Gaussian method (d-5)This article's algorithm 
Figure 1. Comparison of experimental results with three methods 
 
2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018) 
9 
 
It can be seen that in the background frame extraction 
process, the average statistical method extracts the 
background [Fig. 1 (b-1)] and the blurred background is 
affected by the camera shake of the video camera, and the 
effect is worst. 
The parameters of the mixed Gaussian model are set as 
follows: the number of pixel models is 5, the initial variance 
is 30, and the learning rate of model weights is α = 0.005, T 
= 0.7; 
The resulting background [Fig. 1 (c-1)] is clearer, but the 
background extracted at the lower left corner of the screen is 
still somewhat blurred; the corresponding parameter setting 
of this algorithm is: Th = 15, α = 0.2, threshold = 12. The 
background [Fig. 1 (d-1)] obtained by the method proposed 
in this paper is of good quality and closer to the real 
background. 
Finally, after the difference, binary processing, and 
morphological processing, the image is finally obtained. 
The statistical mean method [Fig. 1 (b-5)] is noisy, and the 
extraction result is not clear. The result graph obtained by 
using the Gaussian method [Fig. 1(c-5)] works well, but 
there is still noise. The result graph obtained by using this 
algorithm [Fig. 1(c-5)] has the best effect, the noise is very 
small, the extracted vehicle is clearer and the connectivity is 
better. 
B. Comparison of Algorithm Performance 
The final conclusion of the comparison experiment is 
not only drawn from the experimental results, but also from 
the performance side. Table 1 compares the performance 
ofthe three methods, including the time used for the entire 
inspection process, the memory footprint, and the CPU 
usage. It can be seen that the algorithm proposed in this 
paper compared with the other two methods, it consumes 
less time, consumes less memory, and has a small CPU 
occupancy rate. 
TABLEI Comparison of performance of three modeling methods 
Background 
model 
Time-consuming 
(seconds) 
Average memory 
size (MB) 
Average CPU 
usage 
Statistical 
average 
149 169 65% 
Gaussian 
distribution 
40 99 35% 
Method of 
this article 
25 60 32% 
V. CONCLUSION 
This dissertation focuses on the vehicle detection 
method based on background difference in the intelligent 
traffic field, and proposes a background modeling method 
based on the adaptive difference method between frames. 
Design experiments were compared with commonly 
used averaging methods and Gaussian distribution model 
methodsto compare the background images obtained by the 
three modeling methods. 
At the same time, the running performance of the 
algorithm is analyzed from the aspects of time-consuming, 
memory occupancy and CPU occupancy, which verifies that 
the vehicle detection algorithm proposed in this paper can 
extract the background more accurately. 
The morphological method is used to process the 
differential binary image to eliminate noise, fill in voids, etc. 
to complete the detection step. Experiments prove the 
effectiveness and real-time performance of the algorithm for 
video-based motion vehicle detection. However, the effect 
of this algorithm in complex weather conditions and 
complex road conditions needs to be further improved. 
 
REFERENCES 
[1] Ma Junqiang. Research on the detection and tracking technology of 
sports vehicles based on video [D]. Beijing University of Technology, 
2009. 
[2] Yu jie, Li Xiaojing. Moving object detection based on mean 
background and three framedifference[J]. Journal of Shaanxi 
University of Science and Technology, 2018(1). 
[3] Zhou Mingjiang, Wang Jiwu. Research on the Detection and Tracking 
Method of Moving Vehicles in Video Sequences[J]. Science and 
Technology Economic Market, 2017(4):76-78. 
[4] Fan Wenjie, Zhang Li. An Improved Method of Moving Vehicle 
Detection Based on Background Difference Algorithm [J]. Journal of 
Chengdu University of Information Technology, 2010, 25(4):355-360. 
[5] Huang Lei, Yu Manman. Research on Moving Object Detection 
Based on Background Difference [J]. Software Herald, 
2009(6):187-188. 
[6] Lu Xixin. Detection of moving objects based on Gaussian mixture 
model and three-frame difference method[D]. Xi'an University of 
Technology, 2011. 
[7] Federal Highway Administration, Department of Transportation, 
Highway Performance Monitoring System Reassessment, Final 
Report, FHWA-PL-99-001, 1999: 32-34  
[8] Jean Serra.Introduction to mathematical morphology.Computer 
Vision, Graphicsand Image Processing, 2005,35(7):283-305 
[9] YananakaK.Effects of misdate Smith Predictor on stability in System 
with Time-delay.Automatically, 2011,23(7):787-791. 
[10] Cheung S, Kamath C. Robust background subtraction with 
foreground validation for urban traffic video[J]. EURASIP Journal on 
Applied Signal Processing, 2005,14: 2330- 2340.
 


Not a member yet

Application of Incremental Updating Association Mining Algorithm in Geological Disasters System

Fast Aerial UAV Detection Based on Image Segmentation and HOG-FLD Feature Fusion

The Research and Implementation of 3D Scene Simulation of Camouflage

A New Non-Singular Terminal Sliding Mode Control and Its Application to Chaos Suppression in Interconnected Power System

Semantic Analysis of the Wa Tautology in Subordinate Clauses with Minimal Independence Degree

An Improved Algorithm of the Collision Detection Based on OBB

Advanced Dynamic Autonomous Knowledge Learning Method for Distance Learning

Design and Implementation of Music Recommendation System Based on Hadoop

Research on Vehicle Detection Method Based on Background Modeling

Diagnostic accuracy of sonoelastography in different diseases

4,746

6,096