Computer Vision Solutions

Solutions for the Computer Vision field

The importance of big data utilization is talked about frequently. It is said that extracting the meaning that is hidden in large amounts of data can be used in a variety of ways, for example, to narrow down the next purchasing behavior of a customer, or what leads to guidance programs in the event of a disaster. In the same vein in which large data stores as well as data originating from SNS sites large review sites is currently being utilized, the needs to extract ans use meaningful information from video are also arising.

The installation of surveillance camera systems continue everywhere, howerver, there are customer who wish not to stop at at surveillance, but go beyond the simple storage and use video in more effective ways. Hence, more and more requests for imaginative uses of temporary camera installation requests are being rised.

In our company is ready to assist System Integrators - who coordinate entire customer systems, in various ways regarding the video processing (Computer Vision), and embedded systems fields.

Thinktube Positioning

 TT2

Computer Vision Field

  • Development of image processing algorithms tailored to the specific needs of our customers
  • Prototyping for the verification of the applicability of video processing according to customers requirements
  • Development of utilities for the tuning of processing logic
  • Tooling for the verification of processing logic

Embedded Systems Field

  • Camera requirements · CPU requirements study tailored to processing requirements, platform selection support
  • Benchmarking via prototypes
  • Implementation of video processing algorithms
  • Implementation and development of camera control functions on embedded devices
  • Development of front end interfaces coordinated with backend services

There are several challenges for the camera system and video processing systems, as the processing results are greatly affected by the camera installation location and angle, sunlight and lighting conditions. Also tuning at each site becomes necessary. For the success of a project, it is necessary that customers understand the challenges and considerer how to proceed with the system operation and tuning.

We believe that for the deployment of the system at multiple sites with various different conditions, it is important to foster "video specialist" integrators. We are open to implementing a skills transfer program to face such scenarios.

The Role of Thinktube

imageforCVserviceHP-r

 We present next some examples of solutions developed so far.

Example 1: Parking Area Surveillance

stopdetection-4 50

Example 2: Traffic Counting

traffic count day 50

 

 Example 3: Food Factory - Unsafe Behavior Detection

moving person

 

Example 4: Distributed Image Processing on Embedded Devices

組込み上での処理

 Example 5: Image Processing at Event Hall

cv eventhall customer analysis 01

Installed multiple cameras in a overhead position at the event hall, the video is  transmitted wirelessly, we have a process to capture the dwell time of people in the hall area.

alt

  • Configuration: RMR9000 x 3, IP camera x 1~3
  • Applicability: Retail areas、Foundries、Medical Intitutions、Zoos・Aquariums、Large Parking Lot、Train Stations, etc.