 First of all before we dive in I would like to express my gratitude to all of you for joining to this section in this later hour. And today I would take you on a journey about the intersection of AI technologies and Geospatial World. And today I would like to show you how AI based technologies makes a revolution in interpretation and analysis in valuable special data. Today we will speak about the point clouds. Do you know what is a point clouds? Cool. Our platform this is software as a service solution that is provided different AI models for performing of processing point clouds. Engineers who are working with point clouds definitely know that using of traditional approaches and manual tools for extraction data from the point clouds it's extremely time-consuming challenge and it spends a lot of efforts from the engineers in this. And at the same time today's digital world suggests a lot of AI based technologies and some approaches like using of deep learning, machine learning and I assume almost everyone asks yourself how can I use this in your activities. And our main focus this is classification of the point clouds. While point clouds contains enormous amount of information, raw point cloud this is still like having a book with jumbled letters. Classification at the same time allows you to organize all this data making interpretation of this data and further analysis much easier. For classification our platform suggests currently four standard classifiers which are purposed for different types of surveys. You may use these for extraction objects in different landscapes. For example the most general solution this is standard classifier for airborne laser scans. You may use these for extraction, building, classes, trees, vegetation, vehicles, ground. This classifier is using intensity channel. Apart from the general solutions we also suggest more specified classifiers like classifier for highways. This classifier is able to extract objects that is really one for highways. It may be fences, objects of road infrastructure around the highway, road signs and so on. Apart from the four classifiers that already is available for using at our platform our team currently working on the two more new solutions that is also very general but will be very useful. These solutions will be delivered in the first quarter on the 20-24 and one of them is purr-sport for extraction of objects of road and street infrastructure like road signs, fences, traffic lights, catenary pools and one more solution that we are suggesting for extraction of electric lines. This will be a separate solution. Despite these classifiers can cover a lot of general use cases. There are a lot of cases where you need to solve very very specific tasks. For these cases we can suggest custom solutions. It may be custom classification, future extraction, vectorization and so on. The next example illustrates this very well. In collaboration with Dibinets from Deutsche Bahn we made a specific classifier for extraction of objects around the railways. SMSE, this is platforms, catenary pools, equipment for electric lines, etc. Coming back to time-consuming question, how many time probably you may ask yourself, requires for processing of these point clouds and the next example illustrates this very well. Together with City of Munich we made a classification of scans of whole City of Munich. It was more than high for a hundred squad meters of area. The data set contained 2,000 point clouds with some amount of 84 billions of points and we spent just four days for processing all of this enormous data set. But what about another use cases, not only classification? The next example shows you how AI can be used for extraction of features from the point clouds. At this model that we implemented for performing of three inventories, we performed instant segmentation of the forest and extracted features from each segment. SMSE on the screen, this is type of tree and other parameters. In the end as a result we have map layer which contains information about all the trees inside of forest and information about them. And probably after three days at this exhibition you already know the most challengeable thing in extraction of data in point clouds, this is vectorization. In collaboration with the Autobahn we made AA-based model for extraction of road outlines. At the screen you may see green lines. This is original line that replaced manually. Red lines, this is lines extracted in a fully automatic way using AI. And as you might see there are a lot of cases in processing of point clouds that can be solved using of AI technologies and all of them all of them can be solved potentially at our platform because we provide so wide amount of solutions for different cases. But definitively if you know how works neural networks you have a question where you can get the data for extraction of these features from the point clouds because any AI model firstly requires get training data set. We also can suggest solution for you. Our platform contains some tools for performing of manual classifications. At this demonstration you may see a user can use applied segmentation and selection of different set of points inside of the point cloud and then can assign these selections to different classes. As an outcome of this process you have fully classified point cloud that can be used in your further actions or for training of automatic models. The same is really one for vectorization. Recently we delivered our new features allows you to make vector objects around the point cloud manually. You may draw polylines, place point features or create pelagons. You can combine this manual approach for placement vector objects with classification. This is very efficient way because when you're placing the object on the rail cloud you may skip some details or make mistakes but when you use reclassified cloud for making manual vectorization it increases your efficiency a lot. Recently I would like to have small announcement. We started working on the general solution for extraction of vector data in a fully automated way. It will be AI model purpose for extraction of road outlines and road markings from scans of highways. It will deliver it in the next year and that's why I invite you to follow to our updates and then try. In the end of my presentation I would invite you to sign up at our platform and try to use all these features on your own in your activities and estimate how you can increase your potential using AI, how you can increase your performance. Thank you.