Forest industry app Smart Timber integrates into Segezha Group digital services

2023: Smart Timber Integration into Digital Services

Smart Timber, a product of Computer Vision Systems (part of the LANIT group) for the forest industry, will be introduced into the digital ecosystem of enterprises of the Segezha GroupSegezha Group timber holding (Segezha). LANIT announced this on July 21, 2023. The decision passed preliminary testing. Experts have expanded the number of objects for using the application and added functionality for measuring round timber. Commissioning is scheduled for 2024-2025.

The system for automating the calculation of round forest using algorithmic computer vision and machine learning is being introduced at the Segezha Group enterprises in Karelia and the Vologda region from 2021. According to the annual report of the holding, neural networks helped them to even more effectively control the raw material flow, significantly increase the accuracy of timber measurements, and also reduce the time spent on measuring wood manually.

Since the start of the introduction of Smart Timber, the number of objects for its use has increased. It was carried out integration with other holding systems, the functionality of the solution was significantly expanded. In 2023, it is planned to finalize and test the work algorithms as part of the software and hardware complex and work with different interfaces of interaction, increase the number of data sources in the Smart Timber system and their detail.

The Computer Vision Systems team regularly improves the application and makes it even more convenient for end users. Recent updates include the development of an automatically labeled benchmark. The device is a rail with markers at the ends - this allows you to automatically mark the standard in the image with high accuracy and reduce the measurement time.

Our main goal is to benefit consumers of solutions, enterprises of the timber industry. We work closely with each potential customer to identify issues and propose solutions. The introduction of Smart Timber is often accompanied by organizational changes, and all together this allows you to ensure prompt and transparent accounting of wood with an accuracy that is practically unattainable with manual geometric measurements. We also continuously refine the product based on the results of tests, implementations and feedback from our customers to offer the best solution in the market, said Dmitry Stepanov, CEO of Computer Vision Systems.

2022: Commissioning

The decision to digitalize the calculation of the round forest Smart Timber, developed by the Computer Vision Systems company (part of the LANIT group), has been put into pilot operation at the enterprises of the Russian timber holding Segezha Group. LANIT announced this on April 21, 2022. The system is waiting for the expansion of functionality and scope. During the pilot operation, large batches of round timber were measured using the digital method, the average accuracy of calculations was 99.86%.

Smart Timber facilitates routine processes in the forest industry and improves the accuracy of forest measurements on timber trucks and stacks. This is now confirmed by the successful passing of acceptance tests at the Segezha Group.

Control measurements of 1600 cubic meters of round forest took place on the basis of the Kipelovo timber concern in the Vologda region. They were produced manually, by an independent expert company and by a system. Evaluating the results of the verification measure, experts came to the conclusion that the accuracy of Smart Timber measurements relative to the geometric method is 98.6%, and 97.07% relative to the reference method. The average accuracy of measurements of more than 30 thousand cubic meters of round forest produced during trial operation, relative to the geometric method, amounted to 99.86%.

Also, specialists from Computer Vision Systems and Segezha Group integrated Smart Timber with the Segezha Les corporate wood accounting system at the level of mobile applications and between servers, which increased the economic efficiency of production. When Smart Timber is transferred to pilot operation, the development is waiting for replication, expansion of functionality, zones and methods of application.

"We are very pleased with the successful tests of Smart Timber at Segezha Group - they confirmed the high accuracy of measurements made with the solution and its usability. We see great potential for the use of development in the forest industry and are actively discussing with the enterprise the prospects for further development of the system on mobile and stationary platforms. Cooperation with Segezha Group makes it possible to achieve high performance in the development of the Smart Timber product, said Mikhail Smirnov, acting General Director of Computer Vision Systems.

We have gone a long way from pilots to a fully functional system, proving in practice that digital measurement methods can be applied to measuring roundwood in the field, their accuracy is comparable to conventional geometric methods, and in the future can surpass it. The digital method of measuring roundwood in the medium term could become a new industry standard. Smart Timber algorithms can be used not only on mobile devices, but also as part of stationary software and hardware systems, and complete with augmented reality glasses, added Sergey Merkulov, director of digital transformation at Segezha Group.

2020: Completion of Smart Timber mobile app testing

The Segezha Pulp and Paper Mill Segezha Group in the Republic of Karelia has completed a six-month pilot project to test the Smart Timber mobile application. LANIT announced this on July 17, 2020. It is designed to calculate the coefficient of full timber (efficiency) and the dense volume of round forest at enterprises. This is the development of the Computer Vision Systems company (part of the LANIT group).

As part of the pilot project, it was planned to test the applicability of machine learning and computer vision technologies in a specific application environment. Since part of the data was manually entered at the piloting stage, during the transition to industrial execution, it will be necessary to train the system to work on these reference methods, and not on measurement materials taken directly by employees. In the future, employees, in accordance with special regulations, will only photograph objects, and the program will take over the determination of all calculated values.

Taking into account these factors, according to the results of testing, the accuracy of determining the efficiency was 98.3%, and the dense volume of round forest was 94.5%. The results of the pilot project demonstrated the applicability of the used approach to measuring the volume of round forest and the feasibility of switching to an industrial solution. The project participants positively regard its results - the intended goals were achieved.

{{quote "It has been experimentally established that the service allows you to quickly and accurately calculate the volumes of wood immersed in a timber truck. To calculate the dense volume, it is enough to open the application, take a photo of the stack and enter the required parameters, "says Sergey Merkulov, head of digital transformation at Segezha Group. }}

The mobile application can replace the usual ruler, tape measure, tables and handwritten reports. The neural network independently performs the necessary calculations and provides information on the height, width, length, volume and efficiency of the stack. Photographs of wood, state numbers of timber trucks, geolocation, date and time of the photograph - all data are collected in a single repository with access in 24/7 mode.

The application improves the accuracy of measurements of the volume of incoming wood and minimizes the impact of the human factor on the peer review process. After a certain refinement of algorithms to improve the accuracy of measurements, Smart Timber can also be used for inventory in warehouses.

If a decision is made to implement the software in the Segezha Group, it will be integrated with the SegezhaLes digital dispatching system of the logging enterprise (ASD). It provides online monitoring and control of the entire chain of movement of wood, fuel consumption and fuel and lubricants, and also helps to eliminate non-production losses.

{{quote 'Checking the operation of computer vision algorithms on stationary scan trackers, in mobile applications, 3D reconstruction programs and ML in integration with existing information systems is an important step towards the industrial use of scalable services, - emphasizes Pavel Vakhnin, Member of the Management Board, Vice President for Information Technology and Process Automation Segezha Group. }}

We are pleased that our knowledge and developments have allowed Segezha Group to assess the prospects for using computer vision in their production process. For our company, pilot performance is an important step in the development of our products based on computer vision and their introduction into industrial enterprises of the Russian Federation. Such cooperation shows a good level of industry confidence in the use of high technologies, which will have a positive effect, including on the digitalization of the economy as a whole.