However, it is not always necessary to acquire a vast amount of spectral information per pixel, since there are certain applications where the acquisition of just a few bands is enough. Each of the aforementioned solutions present some advantages and disadvantages over each other, the specific application at hand being the one that tips the balance in favour of one or another.Īs a matter of fact, some of the mentioned technologies are more suitable for acquiring hyperspectral information as they are able to capture hundreds of bands per pixel, while others can just reach up to a dozen of bands, forming what is called a multispectral image. To name some: multi-camera 2D imagers, with as many 2D sensors as wavelengths are captured sequential band systems, recording bands or sets of bands sequentially in time multi-point spectrometers, using a beam splitter to divide the spatial image into sections of which the signal is spread in the spectral domain mosaic filter-on-chip, where each sensor pixel carries a spectral filter built on chip and spatio-spectral cameras, consisting of a continuous filter in front of the 2D panchromatic sensor. Nowadays, there is a wide range of choices when it comes to selecting the camera to utilise, besides the already mentioned technologies. Back in the times where this sensing technology was exclusively installed in satellites or airplanes, whiskbroom and pushbroom imagers were the most recurrent devices to be used, mainly relying on a complex optics system to construct the hypercube. The emergence of new alternatives in the technology for constructing the three dimensional (3D) image cube out of the standard two dimensional (2D) sensor array has certainly paved the way forward. Nonetheless, the incorporation of multi and hyperspectral cameras into drones has only been possible after a great deal of efforts has been invested in the miniaturization of these devices. The latter has been gaining more and more attention as it presents the following attractive advantages: A flexible revisit time, a better spatial resolution, which permits a deeper and more accurate data analysis, and a lower overall solution cost, which has permitted small research groups to start developing their own platforms. It is in the remote sensing area where it has found most of its use, both with sensors mounted onboard satellites as well as onboard unmanned aerial vehicles (UAVs). In the last few decades, multi and hyperspectral imaging have made enormous progress due to the improvements achieved in electronics, computation and software, becoming a very powerful tool for acquiring information relevant to many different fields. Results are presented for the normalized difference vegetation index (NDVI) showing a generated colored map with the captured information. The process was experimentally validated by mounting the camera in a Dji Matrice 600 UAV to uncover vegetation indices in a reduced area of palm trees plantation. In addition, a calibration and characterization methodology has been developed for the camera, allowing not only for quantifying its performance, but also able to characterize other CMOS sensors in the market in order to select the one that best suits the budget and application. The system is compatible with open source hardware permitting one to capture, process, store and/or transmit data if needed. In this work, a low cost and modular solution for a multispectral camera is presented, based on the use of a single panchromatic complementary metal oxide semiconductor (CMOS) sensor combined with a rotating wheel of interchangeable band pass optic filters. These devices make the most sense when combined with unmanned aerial vehicles (UAVs), going a step further and alleviating repetitive routines which could be strenuous if traditional methods were adopted. Different implementations are commercially available from the industry and yet there is quite an interest from the scientific community to spread its use to the majority of society by means of cost effectiveness and ease of use for solutions. Multispectral imaging (MI) techniques are being used very often to identify different properties of nature in several domains, going from precision agriculture to environmental studies, not to mention quality inspection of pharmaceutical production, art restoration, biochemistry, forensic sciences or geology, just to name some.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |