LiDAR & Satellite Incorporated Application(58)
A Studied Case
Overview (1)
Importance of Hyper-spectra and Data Format
Figure 2
Hyper-spectra are one of the most important data which can be widely used in diverse fields. They are capable of identifying substances according to the featured curve of absorption or emission after photons being adsorbed by them. Most functional groups or atoms in molecular structure have their own absorption or emission peaks when they are moving in rotational or vibrational mode. Namely, the absorbed photons cause the movement of electrons at different levels or configurations be altered. The differences of energies are emitted in the form of photons with individual featured frequencies. They can appear in different intervals of wavelengths (wavenumbers).
Therefore, hyper-spectra are very vital and often used in chemical lab.
Now such an analytical device has been implanted into remote sensing technology. Accordingly, a large number of substances on the surface of the Earth are remotely sensed. The situation thereby become very complicated. Such a complexity can be simply classified into two parts.
1. The structure of substance;
2. The structure of storing data.
The complexity of structure of remotely sensed substance is determined by the natural structure and state. There are many discipline associated with studying inorganic substance, and organic substance including biochemical substance. Therefore, a practical topic is how to effectively investigate them by means of remote sensing technology combing with interdisciplinary knowledge.
In order to store the collected data, many different formats have been produced. substantially, most data are stored in binary format. It is a basic machine language in a computer. Hyper-spectra are stored in such a format. Thus, the data can be stored into and read from both memories in a computer and a binary file. The data are stored in order; thus they have sequence numbers while being read into or out of. This is a very important concept especially in manipulating these data.
Importance of Hyper-Spectra and Data Format
A Problem Existing in Dealing with Hyper-Data(1)
Figure 1
As indicated before section: Hyper-spectra sampling (1), hyper-spectra are not allowed to mix with any other data. Once they are handled in this manner, all of important data inclusive of hidden information are completely lost (see Figure 1 and 2). This is caused by the gap between discipline. Most researchers or scientists in spatial science may not have much knowledge about hyper-spectra, which are very sensitive to any signal noises in the process of chemical analysis. Thus, once any disturbing signals are added into, hyper-spectra will completely lose their functionalities to identify substances
Of course, this is a very important but difficult technical task to correctly link the collected spatial hyper-spectra to corresponding spatial points at ground. Because it is very difficult, ones ever tried to merge (mix) them with a multi-band georeferenced imagery. This way could be useful in mixing different multi-band georeferenced imageries. But, this manner used in the digital image processing is not absolutely allowed to use in mixing multi-bands in different imageries or the same imagery with hyper-data.
Now let us to check out a known case. The product (see Figure 1) is produced by such an approach. After test, the problem is easily discovered (see Figure 2), thus hyper-spectra have become completely useless. On the basis of this point, the surveys have thereby lost their meanings.
Accordingly, the scientists and researchers working in the field of spatial geoscience ought to pay a high attention to the method of dealing with collected data.
To fix them, it requires a series of very-high-level specific techniques. In other words, it needs ones to have a lot of interdisciplinary and advanced specific knowledge in order to effectively and properly deal with the existing data, and those to be collected.
Some outcomes of fixed datasets have already been provided in the following subsections as demonstrations:
Some Outcomes of Fixed-Partially-Merged Datasets
