S crop classification [153] or Sea Surface Temperature (SST) retrievals [154]. Even though highly accurate algorithms exist to ML-SA1 TRP Channel detect most clouds, as discussed previously, some papers employ manual masking for higher accuracy [155]. It is also possible to mask deep water, in which coral reef mapping is difficult to obtain. Deep water to be masked can be defined by a criterion including a reflectance threshold over the blue band (45010 nm) [156]. 3.5. Guretolimod Epigenetic Reader Domain Sunglint Removal When operating with water surfaces, such as an ocean or lagoon, sunglint poses a higher threat of altering the top quality of the image, not merely for satellite imagery but for each remote sensing system. Sunglint occurs when the sunlight is reflected around the water surface with an angle comparable to the one the image is being taken with, typically mainly because of waves. As a result, greater solar angles induce extra sunglint; however, they may be also correlated with a greater high quality for bathymetry mapping based on physical analysis strategies [155]. Though this reflectance is usually simply avoided when taking field photos from an airborne automobile (by controlling the time with the day and the direction), it really is tougher to prevent with satellite imagery. It as a result should be removed from the image for greater accuracy of benthic habitat mapping. This could be accomplished, as an example, by a simple linear regression [157]. Some other models can also efficiently tackle this situation [15861]. In accordance with Muslim et al. 2019 [162], essentially the most effective sunglint removal process when mapping coral reef from UAV will be the one described in Lyzenga et al. 2006 [163]. As the procedures compared inside the paper depend on multispectral UAV data, we can think about that the outcome could be correct for satellite data as well. three.six. Geometric Correction Geometric correction consists of georeferencing the satellite image by matching it towards the coordinates of your components on the ground. It permits, as an illustration, removal of spatial distortion from an image or drawing a parallel amongst two diverse sources of data, such as many satellite images, satellite images with other pictures (e.g., aerial), or images mixed with bathymetry inputs (sonar, LiDAR). This step is specifically essential within the case ofRemote Sens. 2021, 13,9 ofsatellite imagery, which can be topic to a sizable quantity of variations for instance angle, radiometry, resolution or acquisition mode [164]. Geometric corrections are necessary to be in a position to use ground-truth control points. These data can take numerous forms, as an example divers’ underwater videos or acoustic measurements from a boat. Even though control points aren’t utilised in each and every study, they may be frequent simply because they allow a high-quality error assessment and/or a more correct coaching set. However, handle points are usually not that quick to obtain for the reason that they demand a field survey, that is not constantly feasible and could be pricey for some remote web-sites. As a result, handle points are certainly not usually utilised. 3.7. Radiometric Correction When working with multi-temporal photos with the same spot, the series of photos is likely to be heterogeneous due to the fact of some noise as an illustration induced by sensors, illumination, solar angle or atmospheric effects. Radiometric correction enables normalization of various pictures to produce them consistent and comparable. Radiometric corrections considerably strengthen accuracy in modify detection and classification algorithms [16568]. Identically to geometric corrections that happen to be only necessary when functioning with ground-truth control poi.