Image Processing Device for Accurate Optical Path Radiance Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices struggle to accurately calculate optical path radiance in wavelength bands where the surface reflectance of dark areas cannot be considered zero, leading to incorrect removal of environment fluctuation components from observation images.
Innovation Solution
An image information processing device that calculates the optical path radiance by extracting radiance values from dark pixels, using irradiance from sunlight and atmosphere information to minimize the cost of difference between intermediate and actual optical path radiance, thereby correcting the observation image for environment fluctuations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the minimum radiance value in the observation image is used as the calculation value of optical path radiance, then the calculation is simple, but the accuracy is insufficient in wavelength bands where surface reflectance of dark areas cannot be considered zero
Solution Approach 1:
The patent performs preliminary classification of wavelength bands into first wavelength bands (where dark area reflectance can be considered zero) and second wavelength bands (where it cannot). This preliminary action enables different calculation approaches to be applied appropriately to each band, improving overall accuracy while maintaining computational efficiency for the first band where the simple minimum value method remains valid.
Solution Approach 2:
The patent changes the calculation parameter from a single method (minimum radiance value) to a dual-method approach based on wavelength band characteristics. For second wavelength bands, it uses alternative parameters such as radiance values from specific pixel regions or interpolated values, adapting the calculation parameters to the specific conditions of each wavelength band to achieve accurate results.
2Measurement precision
If environment fluctuation component is removed from observation image, then accurate ground surface information is obtained, but incorrect removal occurs in wavelength bands with non-zero dark area reflectance
Solution Approach 1:
The patent segments the wavelength bands into different categories (first wavelength bands and second wavelength bands) based on the reflectance characteristics of dark areas. This segmentation allows the environment fluctuation removal process to be applied reliably to each segment with an appropriate calculation method, ensuring that removal is accurate for first bands while avoiding incorrect removal in second bands where the assumption of zero reflectance does not hold.
3Ease of operation
If a single calculation method is used for all wavelength bands, then the processing is uniform and simple, but accuracy varies across different wavelength bands
Solution Approach 1:
The patent applies local quality by using different calculation methods for different wavelength bands based on their specific characteristics. Instead of applying a uniform method everywhere, it tailors the calculation approach to local conditions: using the simple minimum value method for first wavelength bands where it is accurate, and alternative methods for second wavelength bands where additional accuracy measures are needed, thereby optimizing precision for each local wavelength band condition.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate calculation of optical path radiance even in wavelength bands where surface reflectance of dark areas is not zero, improving the precision of environment fluctuation removal and providing effective atmosphere information for weather data.
Implementation Method 1
A pixel value of a pixel, which is included by the image, is an observation value which is outputted by a light receiving element of the image sensor when the light receiving element receives intensity of observation light emitted in an incident direction of the light receiving element
Implementation Method 2
a component of the electromagnetic wave which is scattered by the atmosphere and is inputted into the sensor
Implementation Method 3
absorption by the atmosphere
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In an image representing an observed ground surface area, the present invention calculates an optical-path-brightness with a high degree of accuracy even in a wavelength band in which the surface reflectance of a dark area cannot be considered to be zero. An image information processing device 3 is provided with: a storage unit 31 that associates and stores observation images representing the results of observing electromagnetic waves of a plurality of different wavelength bands reflected from a ground surface, information representing the wavelength bands, and information representing the observation environment at the time that the electromagnetic waves were observed; a first intermediate-optical-path-brightness calculation unit 32 that, for each of the wavelength band, makes the brightness of a dark pixel meeting a brightness standard from among pixels composing an associated observation image a first intermediate-optical-path-brightness; an irradiance calculation unit 33 that calculates the irradiance from sunlight for each of the wavelength band on the basis of the information representing the observation environment; and a final-optical-path-brightness calculation unit 34 that calculates a final-optical-path-brightness for each of the wavelength band on the basis of the irradiance and first intermediate-optical-path-brightness.