How to prepare bathymetric map in QGIS

hello Namaste everyone uh in this video we are going to learn how to prepare mathematic map using qgis and also to calculate the accuracy of the prepared map so let's begin by knowing what is the BC map so basically beety is the study of underwater depth in wason's Lake or other water bodies and bema map visually represent the underwater Topography of that region by showing its depth variation uh through different kinds of color counter lines and a 3D measuring or settings and these maps are very essential for marine navigations or geological uh extractions works and oceanographic resource fisheries management and other environmental monitoring work so for preparing this batc map we requir a different kind of data one of the important data for this is the mathematic survey data which is collected using Technologies such as a sonar system and satellite image image and other type of radar image I think and other are the underwater vehicles are also used for the preparation of those btic data and and these Maps help to reveal underwater features like the sea mountains trenches and re providing inside into the was current sediment and distribution and sea flow geology as well so for the pration of the mathematic map uh there are different kinds of methodologies uh one among them is using the interpolation and the interpolation can be done in the two ways one is the inverse distance waiting method and another one is creating tool so inverse distance waiting method basically assumes that the values at a given location can be estimated by considering the values at the surrounding point with a closure Point receiving more influence than those far away and this method calculates the weight based on the inverse of the distance between the known point and the unknown point which is SC as the point of Interest as well and IDW is relatively simple and computationally less intensive than the cing tool whereas the cing is a geost statical interpolation method this considers both special autocorrelation correlation between point at a different location and the variance of the data set to make the predictions over the another points these involves modeling the spal dependencies structural structure using a variograms which quantify how spal correlation changes with distance and this method then generate the prediction by waiting the known values based on the special structure ging can be more sophisticated and flexible than the inverse distance waiting method it is particularly useful when there is a need to account some special pattern an stropy and when there is a desire to minimize the predictions error by considering the spal correlation structure of the data so we will use qgis version 3.28 in this exercise to prepare the P map and we will use these kinds of data which were collected using sonar in the lake uh which have collected that longitude latitude and the depth point at that location and we will validate uh the interpolated map with the observed Point using a root mean square error which measures how much error there is between the two data set that is me measure data set and interpolated data set now let's begin with the exercise for this we will use qgis 3.2 that I have already discussed so we are now opening the qgis so we can use a new project but I have a similar kind of project so I'm using in this project uh to prepare the map here here so I will prepare a new map as I have talked previously we will be using this data location data and the respective depth data that was collected using Sonar in the lake will be implemented to prepare the Beth metry map in this study so this is Excel CSV file we will importing this into the qgs interface and preparing a interpolation map so let's do it for this we have to import a text elimated layer from here now we will move toward this is the point which I have shown some time before this is the depth data so will open this and we will put the X Val field as a longitude and Y as latitude and we will put Z values as a depth and then add this to the map so uh these are our depth point so where this point Falls we can know by uh adding a base map on this I'm going to add a Google satellite hybrid image here so that we can know the boundary of the lake so this is Cy Lake somewhere in Nepal so this is the betry data and we will using this so first first of all I want to save this layer as a save file layer because it is in the Excel format so I want to save this as a save file format at first so I will save this as a data I will keep the uh coordinate system as this udm Zone 44 which is used for the western part of meal so I will save this so this is our data in sa file format I will remove this uh CSP file from here now we will split these depth data into train data and a test data so that we will use train data for for modeling purpose or the preparing the interpolation map and we will use the test data for the validation purpose so I have already said that we will use the RM value so rmsse value will be calculated using that test data so first of all we will split this data into two sets for that we will use the random selection tool these two and we will input uh dep data and we will consider percentage of feature and we are going to use 30% of the data as a test data so it will select random 30% of the points from these all points as a test data now we will save these points save selected feature as test data so I will save this and remaining data I will [Music] use the invert feature selection tool from here so these are another 70% of the data selected here and I will save them as a train data here now uh these are test and train data both selected both are here into the different layers now we can use train data for the interpolation purpose for this we have to use the processing tool box here and we will pound interation and we we will use uh we have the vector data as a point sa pile so we will use this Saga tool no sorry we will be using this grass Tool uh B do soft do rst no idou so we will use this tool and we will use train data and we will keep this value this is the number of interpolation points means how many number of the neighboring point will be used or will be considered for using the idub method to interpolate the next unknown value this we can put the def uh we can put this values power parameter met uh as default and we can we have to use here the depth data so that it will use the depth column for the interpolation purpose so we can select the extend for that uh I we have to we have to add here another layer which is the lake outside boundary so I will add the leg boundary file from here so there is a leg boundary I open this and add here so we can select extend calculate from layer and there is L layer and we can use the pixel value as for requirement here I'm going to use lake is in a small size I think it is not very much big so I will be using p meter as cell size here and all these things will be kept by default and I will run the code so here we can see the interpolated map repar now we can clip this layer but first we have to we will put this in this color format so this is the batric map interpolate map PR and we will use leg map here we can make this visually something more attractive we will click this feature within this range by doing this so this is the lake area only and this is the bimetric map so these are the training data set these are test dat as I have already said we will will use uh these test data for validation purpose so if we look in the attribute table of this subset there are total of 54 data set here so we will will oh no I think there are 55 data set here which is 30% of the total data set now we will use tool which will insert these restor values as uh we know these are interpolated or something predicted values will be extracted to a new column into this test data set so we will use going to values no values to point I think yes so there is a tool add restor values to point so I will use this tool and here uh test data and I will choose so this is the map interpolated IID out currently prepared so I will this so I will run this picture so here is the result we can see the attribute table there is the depth data which is measure data and interpolated data which is the predicted data by our ID interpolation method so we will use interpolative depth and this sometime this is known as the predicted and this is obsorb data and we will use formula here mentioned R minus obser square summ and there is divid by n and the whole square root of these value will provide us the root Min Square value which measures the error of this prediction so for that there are two techniques we can select all the data by this method we can copy selected layer from here we can go to Excel and we can calculate rmsc value but I'm going to calculate these layers here so for this I will be going to the rest calculator here I will create a new field as r m s E1 because it is not uh the whole process cannot be completed a single step so I will use add this as one so this will be decimal point I will use Precision up to four point so the formula is predicted minus so so here we can see predicted minus observed squ and summation so we will use up to this step in this column so it is Open Bracket so I interpolate minus so I will use this minus here minus depth which is observe value and is sire this P which means this part of the formula after that we are going to sum these all values and divide by the total number of the observation so this will provide us this value here you can see the one of the result preview result here so it will calculate all the values sorry for this I have select single layer so it provided only a single layer value so I will discard this and I will deselect all layers from here so I will go here again do the same thing RMS you1 I will use decimal real number precis of to four so I will do R interpolate minus up to dep then square of this this is working for now so in the next we have to sum these all so sumide by N I will do sumide by n so I will put this name to this is just for example so you can use uh these operation into the Excel seat as well so this sum means I'm going we have calculated B minus Square whole Square uh and we are now doing the summation of these these values Su of r m s E1 so this is .61 so if we do divide by y then we can put is divide by we have already known uh there are total value 55 here so it is 55 so uh this is 0.66 and if we do sqrd of hold the values then we can get the r oh sorry again I just forget to keep sorry for this we will do it again because I forgot something here SF so I just forget to keep this as a deal number so I will put this and we will keep SQ are the square root of square root of sum of r m E1 divided by 55 and this this is our final AR this is so this is very much accurate so we can check this into the Google r m e a this means so rmsc value between 0.2 and 0. shows the model can relatively predict the data accurately so you can see it is good so this is all we have prepared map as well as calculated the accuracy of the model so after this calculation I have here already prepared some of the maps out of this uh here so in this layer I have prepared the counter map by using the interpolated layer uh these are the counter lines and and poly counter polygons after that and we can also make these layers uh into the 3D format and we can show a 3D visualization of the layers like this which I think is very effective to present as well so this is the layer yeah you can see this this is final map so I can make another video so how to prepare uh the 3D image of the model so if you are interested to learn how to make this 3D image out of the interpolation map so please comment here and I will make another video of that as well so thank you for now uh I will Stu here

Channel: Virtual tech GIS and RS

Views: 2,087

Keywords: Interpolation, Bathymetry, QGIS, RMSE in QGIS, Bathymetry in QGIS

Id: dGDr3BrmmNw

Length: 23min 53sec (1433 seconds)

Published: Sun Feb 11 2024