Technology: Computation Framework
Scorch Computation Framework is aimed to simplify the creation of complex AI systems
  • Combination:
  • Combines various machine learning modules, as well as auxiliary modules into a single ecosystem
  • Integration:
  • Provides integration of input and output data with business formats


Devices: consumer and industrial drones, consumer cameras, surveillance cameras
How it works:
  • Modules are offered as embedded machine vision application with basic configuration
  • The configuration is managed with update files or new vision skills from the Cloud Hub so embedded machine vision meets the needs of a customer
  • The video stream generated by the camera of a device is grabbed and analyzed by the modules and according to their settings
Each module’s settings can have triggers so, when something is visually recognized, the module sends a signal to a device. The triggerable events and situations are:
  • A certain object, type or group of objects have been detected and localized in a video
  • A human or group of people have been recognized in a video
  • Specific action or actions produced by a human, group of people or as a result of two or more objects interaction have been recognized in a video
According to the signals provided, a device understands what to do next, e.g. approach (drones), start videoing, create an image, contact operator, etc.
The modules can be connected to the Cloud Hub so:
  • New updates for the selected scenes and activities are available to a device
  • New and unknown scenes and activities are sent to the Cloud Hub so they are learned
Some Scorch module commercial and consumer drone use cases
Large construction and other large industrial areas surveys
Drones with Scorch modules can assess the effectiveness of construction site processes by identifying various types of workers behavior and machinery movements. The data obtained by the modules can be extrapolated into important key factors.
Surveying is effective for large construction sites and large objects under construction, mineral fields and quarry development, railroad repairs, agriculture activities as well as for hard-to-reach places such as the ones during a bridge construction.
Some key performance metrics can be made observable such as a number of workers, their level of activity, statistics of their movements and time of their being at a certain location.
That will help to improve business processes, personnel and machinery activity and positioning for construction, agriculture, mining and quarrying businesses.
Filming and entertainment drones
One of the use cases with search algorithm:
There are multiple surfers of various skills being visible by a drone from a distance.
This drone is able to notice the ripping* surfer from a distance by estimating the probability of a sought skill level and recognizing a distinction between the surfers’ actions, which look similar.
Based on that, the drone would be approaching a sought-action surfer with having its camera keep zooming in on him/her until the beDer recognition rate becomes possible.
If the skill level of the ripping as such does not meet the criteria, the drone reverts to its previous position and continues ‘eyeing out’ for surfers with the sought skills.
Once the sought surfer has been detected, the drone would have its camera record him/her.
* A surfer who rips is one who is seriously skilled, and often shows this off in the view of other surfers or spectators.
Drone surveys for Agriculture
Scorch-modules equipped drones can autonomously survey and detect minor changes in crop performances, match received data with weather conditions and fertilization schedule. The data collected in the Cloud Hub becomes available for the analysis and building of a predictive model for deeper knowledge.
Scorch modules can autonomously collect accurate data for particular fields and seeds, seasons and weather/ humidity/soil so enabling to create a very detailed model of local agriculture. Scorch modules are able to signal about the very first signs of changes in crop performance.
Distant pipelines surveys
Scorch modules for the autonomous distant surveying of oil and gas pipelines so a drone is able to make decisions based on the data recognized by the modules embedded.
A Scorch module is able to detect gas leaks based on the the analysis of a IR video. By detecting the leak and receiving an analysis result about how bad the leak is, the drone would make a respective decision C e.g. send a message and go further, or stay put and monitor the dynamics or repair the leak.
A module-equipped drone is able to detect more than one leak at a time and locate the worst one that must be handled first.
Every use of such drone will save utility costs, improve safety and increase the density of a survey.
Scorch module surveillance camera
use cases
  • Misconduct and riots
  • Analysis of crowding
  • Objects/items left and unattended, detection of situations when an item was picked up by another person
  • Suspicious behavior analysis
  • Discerning between humans and animals
  • People counter and maximum crowd capacity reach alert - for providing better services or preventing crowd disasters
  • Fires, pipeline breaches, other accidents and emergencies
Scorch module consumer video camera
use cases
  • Video auto-recording of the events/ objects of interest or auto-marking of them for editing purposes.
  • Still image auto-taking of the events/ objects of interest and auto-tagging of them in real time.
  • Auto-focusing on the objects and events of interest with bokeh added to the background.