A topographic or thematic map and an object database: type, dimensions and changes over a period at the agreed scale.
Search the site
What the service is used for
Cartographic materials
Creation of cartographic materials and topographic interpretation for maps.
Inventory of terrain changes
Recording changes of objects and the territory over a selected period.
Medium-scale updatable maps
Extensive spatial coverage when creating a medium-scale updatable map.
Agriculture, forest and urban infrastructure
Composition and changes of fields, forests and urban infrastructure from imagery.
Geology, water and ecology
Geological structures, condition of water bodies and changes in the natural environment.
Did not find your task?
Describe the territory, dates and objects for interpretation — we will assess the remote sensing archive and the scope of work.
What you will receive as a result of the work
01
Topographic or thematic map
A map with interpreted objects plotted at the requested scale and for detection tasks of features on the satellite image.
Objects are georeferenced to the terrain. Scale and composition depend on the environment and the terms of reference: general (geographic) or specialized (thematic) interpretation.
What you get
- a topographic or thematic map with interpreted objects
- georeferencing of objects to the terrain
- composition by scale and detection tasks
Format, scale and coordinate system are fixed in the contract and the terms of reference.
02
Object database
Interpreted database: a set of objects from satellite images with methods of primary data pre-processing.
The set is used for testing, comparison and training. The terms of reference fix the required number of reference objects to achieve accuracy.
What you get
- an interpreted object database
- a set with methods of primary data pre-processing
- requirements for the number of reference objects
Attribute composition and database format are agreed in the contract.
03
Algorithms and machine learning
Training of existing approaches and algorithms for automatic object detection on satellite images using neural networks and machine learning.
Deep machine learning provides an accurate and efficient solution for object details on satellite images — as a separate order. Computational costs can be reduced by using low-resolution images.
What you get
- training of existing approaches and algorithms on various satellite images
- algorithms for automatic object detection (as a separate order)
- methods to reduce computational costs on low-resolution images
Neural-network algorithms are included in the scope of work only as a separate order.
04
Final report
Interpretation results are recorded in graphical, digital or textual form and delivered together with the final report.
After processing and interpretation of remote sensing materials, the result is checked and a report is prepared. The composition depends on the task: map, database, object characteristics.
What you get
- a verified interpretation result
- a final report
- materials in graphical, digital or textual form
Delivery form and report composition are agreed in the contract.
How the work goes
Cost and timeline
The cost depends on the plot area, the number and quality of images, terrain and task complexity, seasonality, advance-payment size and computational-capacity requirements. The timeline depends on the area of the territory and the requirements for the result.
- consultation — free of charge
- interpretation of remote sensing materials — the cost is calculated individually and depends on the complexity and volume of work
- work of technical specialists and expert(s) — from 100 000 RUB; TOTAL — from 100 000 RUB (table and order card)
- in the order steps: study from 100 000 RUB; timeline from 20 working days from the date the advance payment is received; payment by bank transfer only
- timeline on the card — from 5 days; in the timeline block — from 10 working days from the date the advance payment is received, calculated individually for each client
Order cost — from 100 000 ₽
Timeline — from 10 working days
Consultation — free of charge
What data are needed to quote interpretation
To assess feasibility, cost and timeline, provide:
- location of the study object (coordinates)
- questions to be answered from the images
- dates for which interpretation is required
- the task, area size, terrain character and product requirements
- if there are requirements for primary images: georeferencing accuracy, spectral range, off-nadir angle, radiometric resolution, cloud cover and snow cover
If survey parameters are unknown, describe the task — requirements for satellite image processing are agreed with the contractor.
Describe the task in simple terms — we will assess the archive and the scope of work.
Why Innoter
Prompt access to archives
Suitable archive imagery can be obtained faster when the required area and date are already available from operators.
No aviation clearances
A new satellite survey does not require the flight clearances typical of aerial surveys.
Large and remote areas
A single satellite pass covers large areas and makes it possible to work in hard-to-reach regions.
Direct agreements with operators
Distribution agreements help select archive imagery and order new surveys from different providers.
Software and server infrastructure
Modern software and computing capacity for quality control and processing of large data volumes.
Experienced specialist team
Years of experience on complex projects and specialists in cartography, photogrammetry and remote sensing.
Interpretation types, sensors and source data
The composition depends on the task, survey resolution, archive or new survey, and object classification requirements.
1
Remote sensing data interpretation is the process of recognizing objects and territories, their properties, and interrelations based on their images on a snapshot. It can be field-based or office-based. Office work is divided into visual (by eye) and automated.
Automated (machine) interpretation is performed in software suites using algorithms and boils down to classification mechanisms. Supervised classification: Minimum Distance Method, Spectral Angle Method, Mahalanobis Distance Method. Unsupervised: ISODATA, K-Means Method.
Depending on the tasks, interpretation is classified as general (comprehensive, geographic) and specialized (thematic, specific). The process includes preliminary and main stages: data processing, brightness normalization, and creation of mosaic coverages. Results are recorded in graphical, digital or textual form.
2
Stage I. Detection — the initial, lowest level: searching the image for areas where terrain objects are most likely depicted. The operator notes: “There is something here.”
Stage II. Recognition — the middle level: determining the nature of the detected objects. The result is to recognize, fail to recognize, or recognize incorrectly.
Stage III. Determination of characteristics of identified objects — the highest level: analysis of quantitative and qualitative characteristics, condition and significance. Parameters are measured from the photo image: geometric dimensions, parallaxes, densities. This is how forest composition, soil character, road-surface material, linear dimensions and distances between objects are established.
3
High resolution 0.15–1.0 m (spacecraft images 0.3–1.0 m): Maxar, Planet, AirBus Defence, SuperView-1 (GaoJing-1), «Jilin» series KF-01B, L-SAR 01A and L-SAR 01B, «Ресурс-П». Tasks: urban development, buildings, structures, land and forest plots, streets, roads, utilities, restricted zones. Scale 1:2000–1:10 000. For urban infrastructure typically 0.3–1.0 m; local sites — aerial or UAV 3–25 cm.
Medium resolution 2–7 m: Spot 6/7, ALOS, Kompsat, «Канопус» series. Scale 1:25 000–1:200 000. Water bodies, agricultural land, extensive areas. For thematic mapping, 1.5–5.0 m sensors are used more often.
Resolution 15–60 m: Landsat and Aster — up to 16 filters of the optical and IR ranges, including thermal. Images of 30–100 m are useful for thematic mapping over large areas. Combined processing can, if needed, increase spatial detail when the data are of good quality, for example Landsat.
4
A comprehensive approach to remote sensing interpretation relies on specialist experience and software and hardware. Combining panchromatic and multispectral data in a multi-purpose index improves spectral quality and spatial detail compared with classical approaches.
Index values are influenced by vegetation species composition, canopy density and condition, and to a lesser extent by exposure and surface slope. An example is NDVI. With a UAV, object texture can be extracted faster than by fieldwork, which does not cover height characteristics, while structural changes may take years to study.
5
Interpretation is affected by differences between object classes, several objects in one frame without details, a wide range of sizes, illumination, dense and noisy sensor background, object motion and change over time, and an unusual viewing angle.
Two professional problems: accuracy (a wide range of intra-class variation) and efficiency (a huge number of categories). Intra-class variance: the visualization state and an internal factor — position, colour, texture, material reflectance, surroundings and lighting. Distortion, occlusion, scale, blur, background clutter and shadowing also affect the result.
Therefore requirements for the primary sensor survey (optics, IR, radar) matter: season, solstice, viewing angle, spacecraft time and coordinates, cloud cover, weather conditions, spatial resolution.
6
Resolution of the Government of the Russian Federation of 10 June 2005 No. 370 “On approval of the Regulation on planning of space surveys, reception, processing, storage and dissemination of Earth remote sensing data from civil spacecraft of high (less than 2 metres) resolution”; Resolution of 7 July 2015 No. 682 “On the powers of federal executive authorities in the field of using the results of space activity”; Roscosmos orders and Government of the Russian Federation resolutions on the federal fund of space-borne remote sensing data.
As of 2022-03-30, 36 remote sensing standards have been approved by Rosstandart orders. Internationally: Convention on the Transfer and Use of Data of Remote Sensing of the Earth from Outer Space (Moscow, 19 May 1978); Principles Relating to Remote Sensing of the Earth from Outer Space (UNGA resolution 41/65 of 3 December 1986).
Works are performed in accordance with SNIP, GOST and SP, advanced methods and modern software.
Frequently asked questions
Answers to key questions about service parameters, timelines, deliverable formats and workflow. If you did not find what you need — contact us and we will help.
Primary snapshots:
- acceptable accuracy of spatial geolocation of images;
- spectral range of shooting: panchromatic, multispectral (color), color with near-infrared channel, color with short-wave near-infrared channel;
- the angle of deviation of the image from the nadir, for example, up to 20 degrees or up to 10 degrees;
- radiometric resolution of images, for example, 8 bits /16 bits (the higher the resolution, the more brightness gradations can be seen in the image);
- the percentage of clouds in the image, for example, is no more than 20%;
- snow cover in the images (20% of the area).
If space image processing is required, the requirements are discussed with the Contractor.
For high spatial resolution (0.15 – 1.0 meter), images from spacecraft (0.3 – 1.0 meter) Maxar, Planet, AirBus Defence, Chinese: SuperView-1 (GaoJing-1), «Jilin» series - KF-01B, L-SAR 01A and L-SAR 01B, Russian «Ресурс-П».
Their survey addresses detailed interpretation of built-up urban areas, buildings, structures, land and forest plots (details of vegetation change), streets, squares, avenues, roads, utilities and restricted zones and other lands at cartographic scales 1:2000–1:10 000.
For medium resolution (2–7 m), high-quality imagery from Spot 6/7 spacecraft, the Japanese ALOS series, the South Korean Kompsat series, and the relatively satisfactory quality of the Russian «Канопус» series. Scale 1:25 000–1:200 000.
Such imagery helps interpret water bodies, agricultural land and land use, and sufficiently extensive areas.
Finally, Landsat and Aster imagery at 15–60 m resolution: thematic interpretation of vegetation and land is an order of magnitude stronger than the sensors above, because they have up to 16 optical and IR filters, including thermal.
Analytical combined processing allows you to pull out, if necessary, and high spatial resolution, but with high-quality data, for example, the Landsat spacecraft.
Yes, of course, if you need to increase the detail of the object under study, pull out structural features from the image of the object, create 3D models, an overview of 3600, a multimedia product, get an oblique shooting and texture of the image.
Organising such monitoring consists of multi-stage comparative processing of satellite images of the same area over a defined period (5–10 years), with interpretation of changes in urban development, the road network, the electrical network, telecommunications infrastructure, soil, land, vegetation and water resources.
The most modern methods of interpretation, tested in practice, are the methods of so-called “artificial intelligence” (machine reading), based on the selection of objects from neural networks, automatically comparing prepared standards, installed objects, with objects in the image. However, not everything is so simple.
There is a satellite-image dataset for object detection using convolutional-neural-network frameworks such as Faster R-CNN (faster region-based convolutional neural network), YOLO (you only look once), SSD (single-shot detector) and SIMRDWN (multiscale fast detection of satellite images with windowed networks). These approaches have also been analysed for accuracy and speed using a developed dataset of objects on satellite images.
The result of thematic interpretation, classification and spatial analysis performed by an expert using Earth remote sensing and geoinformation technologies is a thematic cartographic product (a thematic map, a thematic database). Its creation can be called the final stage in using satellite survey materials. In mapping, expert specialists from different industries identify the objects and parameters the client needs based on optimal combinations of satellite imagery. The result of thematic mapping may be a map of any thematic focus, or a database with visualization in GIS.
Satellite and aerial images of any spatial resolution are suitable for thematic interpretation of satellite imagery, depending on the final task
Need satellite imagery or help choosing the right data?
Describe your task, area of interest, and desired outcome. We will check data
availability, assess new acquisition options, and propose the optimal solution in terms of timelines, cost,
and quality.
Discuss the project
clients who trust us
25+ years in the geodata market
