The concept of spatial resolution

In the earliest aerial photography remote sensing, spatial resolution is the most important technical index, which directly affects the visual recognition of ground objects. Spatial resolution is defined as the line pairs that can identify the bright and dark stripes on the chopping block in a unit distance, and the unit is the line pairs per millimeter (lp/mm), that is, the smallest ground distance that can be identified or the size of the smallest target object when converted into an actual object in proportion. For aerial black-and-white photography with the scale of 1: 10000, the resolution of RMK top-level aerial camera lens is 94lp/mm. According to the formula of ground resolution: Rg=m/AWAR(m is the denominator of aerial photography scale, AWAR is the comprehensive resolution of aerial photography, such as 94lp/mm), the ground resolution of this scale is calculated to be 65438+.

In multispectral and imaging spectral optical scanning technology, spatial resolution refers to the measurement of the minimum angle or linear distance between two identifiable targets, which is generally expressed by the instantaneous field of view angle IFOV of the sensor, and the unit is milliradians (mrad). Ground resolution The ground area covered by IFOV is usually called pixel size or dimension. Calculation formula:

IFOV = high× signal-to-noise ratio (5-5- 1)

Where h is the height, s is the side length of the detector, and f is the focal length of the system.

According to this formula, the spatial resolution of remote sensing image is affected by the height of remote sensing platform, the size of detector and the focal length of optical scanning system. From the above definition, we can know several manifestations of spatial resolution (Zhao, 2003):

(1) line pair. Mainly for aerial photography system, the minimum unit of the image is determined by the logarithm of the line contained in the 1 mm interval. Generally speaking, the higher the resolution of aerial photography lens, the clearer the aerial photography image.

(2) pixels. Refers to the size of the ground area corresponding to a single pixel, in m or km. For example, the pixel area of Quickbird satellite with high spatial resolution in the United States is 0.6 1m×0.6 1m, its spatial resolution is 0.6 1m, and the spatial resolution of TM sensor is 30m. The size of the pixel is determined by the height of the platform and the instantaneous field of view. For the imaging spectral image scanned by optical machine, the ground resolution changes with the change of image point position, and it is the highest at the sub-satellite point, and the vertical resolution and horizontal resolution are equal. The ground resolution of other positions gradually decreases from the middle to both sides, and the vertical resolution and horizontal resolution are not equal.

(3) instantaneous field of view (IFOV). IFOV is an angel. The smaller the IFOV value, the smaller the minimum resolvable unit and the higher the spatial resolution.

However, in any given IFOV, there is often more than one type of ground coverage, and the spatial information it detects is a comprehensive signal. Therefore, pure pixels and mixed pixels often appear, and the degree of mixing depends on the size of IFOV and the spatial complexity of ground objects.

Generally speaking, the higher the resolution of the sensor system, the stronger the ability to identify objects. But in fact, the resolution of each target on the ground in the image is not completely determined by the specific value of spatial resolution, but is related to its shape, size and relative difference with the brightness and structure of the surrounding objects. For example, the MSS spatial resolution of LandSat is only 80m, but the railways with a width of only 15~20m, and even the expressways with a width of only 15 ~ 20m, are often clearly distinguishable when passing through areas with monotonous background spectrum or large road spectrum differences, such as deserts, waters, grasslands and crops, which are caused by this unique shape and relatively single background value. It can be seen that spatial resolution only indicates the visibility of image details, and the real recognition effect depends on the complexity of environmental background and other factors. Therefore, for regions with monotonous and uniform feature types, the requirements for spatial resolution are relatively low; For the complex area of the ground object and the micro-morphological structure of the ground object to be detected, it is necessary to improve the spatial resolution of the sensor as much as possible in order to extract the detailed information of the ground object. Experience shows (Jensen John R., 1996) that the spatial resolution of the sensor system should usually be less than 0.5 times the minimum diameter of the detection target. For example, if an oak tree in a park is identified, the minimum acceptable spatial resolution should be half the diameter of the smallest oak tree crown. However, if the spectral response difference between oak trees and background features is small, the spatial resolution recommended by empirical values is also problematic (Zhao, 2003).