+7 8202 20-21-17
Mon-Fri 9:00AM – 6:00PM (MSK)


For 20 years, Mallenom Systems specialists have developed and successfully implemented dozens of machine vision systems. In our solutions, we use both our own unique developments in the field of image analysis, as well as proven algorithms from the world leader - Cognex. For tasks where recognition is required to be performed in arbitrary, changing external conditions, we use artificial neural networks.

Find below several of our OCR (Optical Character Recognition) projects.

Reading the marking on fuel rod arrays

Alphanumeric markings on fuel rod arrays must be read with 100% accuracy. To do this, Mallenom Systems developed special algorithms based on machine learning. The recognition model was trained on images of real monitor objects.

Both during tests and in operation, the required recognition accuracy was confirmed.

Reading marking from blasting caps

Marking 4 mm in size is applied to the 8 mm diameter blasting caps.

To reliably read the marking, Mallenom Systems developed a system that uses Cognex VisionPro library and software for the operator controlling the process.

On one computer, parallel marking of 6 objects is recognized.

Recognition of embossed text labels of industrial blanks

Recognition of digital marking on the ends of metal slabs for heavy machinery is often hampered by presence of rust that occurs during their transportation and long-term storage. As a result, the characters become low-contrast, which significantly complicates their reading.

Under these conditions, reliable detection and character recognition were made possible through the use of a 3D scanner.

Reading railcar numbers

Identification of railcars is performed by recognizing the registration numbers located on the sides and chassis. This operation is performed in different weather conditions, with different optical schemes on all types of wagons and locomotives with different degrees of dirtiness.

Mallenom Systems developed the ARSCIS system, that was implemented at more than 30 enterprises in Russia and the CIS countries. At the heart of ARSCIS are recognition algorithms based on neural networks, which make it possible to read even the most illegible numbers with high accuracy. The highest reading accuracy is achieved through the use of the algorithm that recognizes and compares all 4 duplicate numbers on the railcar (2 on the sides, 2 on the underframe) on a series of shots from 4 video cameras.

The software allows recognizing the numbers even when the railcar stops or changes direction.

Reading marking from steel pipes

Mallenom Systems developed the software to read marking from steel pipes. Since the marking is applied on pipes with poor quality, and the pipe itself rolls over in the control zone, recognition is performed on a series of frames. For this purpose, the results of analysis of individual frames obtained from Cognex VisionPro library fall into the analytical software module of the upper level, where the resulting decision on marking is made.

The accuracy of marking recognition by frame series is 99.8%.

For additional information, please contact:

Tsareva Ekaterina, Project Manager
e-mail: tsareva@mallenom.ru
phone: +7 (8202) 20-16-39, cell: +7-921-251-62-96