A Practical Method of 
Remotely Sensed Digital Image's Resolving Power 
Izumi Kamiya 
Geographical Survey Institute of Japan 
Kitasato-1, Tsukuba-shi, Ibaraki-ken 
Japan 
Commission I 
1. Abstraet 
A method which decides resolving power of digital remote 
sensing sensor in field experiment was developed. It was 
applied to MOS-1 airborne verification experiment and predicted 
a resolving power of MESSR on MOS-1 satellite. : 
An outline of this method is followins. An airphoto sjienarl, 
refrectancs of which is like step. function, is.laid slightly 
obliquely (angle &) to perpendicular to flying direction and 
observed (Fig. 1). It is equivalentobouwobsenving«a "object, 
radiance of which changes like step functions: in’ various 
phases. A calculation method of PSF and MTF from the data was 
developed. 
2. MOS-1 and MESSR 
  
NASDA(National Space Development Agency of Japan) launched MOS- 
1(Marine Observation Satellite. LY dan 92 Le BELO Sr LIS 
launching, NASDA conducted MOS-1 airborne verification experi- 
ment in order to test sensor equipment. GSI(Geographical Survey 
Institute of Japan) took part in the experiment, and evaluated 
resolving power of MESSR(Multi-spectral Electric Self-scanning 
Radiometer). It is & visible and infrared image sensor on the 
MOS-1. Table 1 is selected technical data of the MOS-1 and the 
MESSR. 
The MESSR employed linear CCD(Charge Coupled Device) sensor 
like SPOT HRV. One of the most significant difference between 
linear CCD sensor and seanner(ex. LANDSAT MSS, TM) is follow- 
ing. Seanner samples incident light in IFOV(Instantaneous Field 
of View) in a moment. But CCD accumuliatesiincddent:lightitin 
IFOV while satellite moves corresponding length between two 
consecutive lines. So, the area that a detector observes 
changes during the accumulation. It contributes: to lower 
resolving power in flying direction, but no ‘contribution in 
perpendicular to the flying direction. 
3. General idea of Ihe method 
In order to obtain PSF(Point Spread Funetion) in certain direc- 
tion, it is enough to observe airphoto.signal, refrectance..of 
which changes like step function in the direction, in various 
locations of sensor relative to the edge. But it is impossible 
to design a airplane or satellite based experiment to fix a 
pixel to certain location. Also it' is very diffieult to observe 
the Locations of pixeles in sub-pixel order. So, statistical 
processing of edge observation data, in which every locations 
were thought to appear in same probability, was used for field 
64