The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B3b. Beijing 2008 
difficult to obtain in public because of its great military and 
political significance. Also, Escort Micro-Satellite, presented by 
American AeroAstro Company, owns the function of on-orbit 
checking and proximity except XSS series. In China, many 
experts have researched the motion analysis and controlling for 
non-cooperative target toward the relative orbit deeply, but the 
researches about the relative navigation surveying are very few 
[5-11]. Based on single CCD optical camera, Zhang el al. 
realized the iterative algorithm on the relative location and 
attitude parameters surveying for non-cooperative spacecraft 
according to known structure model information of target 
spacecraft, which affords the paper a salutary lesson [12]. 
2. FRAMEWORK OF THE RELATIVE NAVIGATION 
SYSTEM 
Different from the relative navigation system for cooperative 
target, the motion rules of the relative orbit between the non- 
cooperative target (called “target” for short at the following 
section) and tracking spacecraft must be considered in the 
relative navigation system for the target. And the entity model 
for the target can be simplified as the feature point model which 
can be observed by way of digital image processing utilizing 
the structure characters of the target. Then the task of the 
relative navigation will be realized using digital 
photogrammetry method based on “point surveying”, which 
includes the determination for the relative location, attitude, 
velocity and angular velocity between the target spacecraft and 
the tracing spacecraft. 
2.1 Analysis and Design of the Relative Orbit 
Because of the randomicity of the target spacecraft’s orbit, the 
relative motion features between the tracking spacecraft and the 
target can be only analyzed in a general way to obtain a relative 
motion equation based on general elliptical motion and solve 
the analytic expression of the equation owning arbitrary initial 
condition, which will provide theoretical basis for researching 
relative motion features and designing motion construction. 
And the relative orbit where the tracking spacecraft follows the 
track of the target will be analyzed and designed under 
integrating different constraints which include the orientation 
for the tracking spacecraft toward the earth and the observation 
scope of the satellite-borne CCD. Then the motion orbit of the 
tracking spacecraft will be designed based on the deep analysis 
of the relative motion equation to realize the tracking survey, 
which can provide some priori parameters and initial values for 
relative navigation. 
Because the target orbit is an approaching circular orbit, the 
relative orbit motion at the proximity stage can be described by 
the C-W Equation which is also called as the Hill Equation [13]: 
0 2 +- 
20 
1 + ecos# 
f _ 
y + Ox + 2 Ox — 0 2 
x-0y-20'y - 0 
2 A 
z + - 
0 2 
(f_ 
1 + ecos 0 
■z = 0 
(2) 
1 -i-ecos 0 
where 0 is the true anomaly of the tracking spacecraft in the 
equation. 
According to the C-W Equation and T-H Equation, the 
constructions of the relative orbit will be discussed and the 
CCD’s tracking angles, scopes and the resolution constraints of 
the relative navigation at every construction will be analyzed 
too. The process is shown in Figure 1. 
Figure 1. Analysis and design of the relative orbit 
2.2 Quick Image Matching Base on Features 
The image matching of spatial target will meet a greater 
challenge in comparison with traditional remote sensing image 
matching. These reasons are: 1) the speed of image matching 
need be raised in order to ensure the response frequency of 
relative navigation system. Because traditional image matching 
algorithm is rather time-consuming, a new matching mechanism 
will be adopted which will utilize the interior relevant features 
of sequence images fully in this paper to make the matching 
rate as fast as possible on the premise of keeping the matching 
accuracy; 2) the image of spatial target has obvious difference 
with geo-image of remote sensing. Under the limit of sensor’s 
view-field and lens’s field-depth, the target image will be 
relatively small and the textures will be rather unitary when the 
tracking distance is quite far. Moreover, the target background 
will be rather complex under the influence of space background 
and other space environments so as to cause a lot of trouble to 
the image matching. Therefore, some effective pre-processing 
methods need be used in order to remove pseudo image and 
background noise and extract useful and reliable target 
information to match. The process is shown in Figure 2.