X5 Retail Group mastered computer vision

2019: Implementation of video analytics in supermarkets of Moscow and the Moscow region

On September 30, 2019 the X5 Retail Group company announced TAdviser start of the pilot on video analytics of shelves and zones of fruit and vegetables in the retail chain stores. The solution allows in real time 24/7 to monitor existence of goods and to timely inform personnel on need of replenishment of shelves.

The pilot is carried out in 38 Pyaterochka shops and 6 Perekryostok supermarkets of Moscow and the Moscow region. After its carrying out and achievement of indicators of cost efficiency the decision on deployment of the project on other shops of networks will be made.

The system of video analytics is completely developed in directorate of Big Data X5. The solution is customized under each network, its assortment matrix and the existing features of work of personnel.

The technology works in a zone of fruit and vegetables and also in the most wrapped categories (for example, dairy). Computer vision in real time will recognize objects of different classes and also fixes their absence. In case of a deviation from target parameters there is an automatic notification of personnel of shop. The management of network obtains the aggregated information on work of shops in a report type. All data remain for the purposes of the subsequent analytics.

In department of fruit and vegetables the artificial intelligence is capable not only to analyze emptiness in general, but also to define existence or lack of separate goods on shelves and in boxes and also to fix violations of appearance of a zone. Due to use of modern analytical algorithms and the choice of the optimal locations of cameras a system allows to minimize influence of the factors disturbing the overview of cameras. The solution in the automatic mode fixes approaches of employees to the shelf, including consolidates information for determination of the optimal time expenditure necessary for maintenance of goods availability close to 100%. It promotes effective planning of load of personnel.

The corner and most difficult question for successful implementation of video analytics in shops is a creation of the "correct" business processes. It is long iterative work in close cooperation with networks for this reason in X5 the decision on creation of in-house of examination on video analytics was made. Any more there are no doubts that emptiness on shelves and also appearance of a zone of fruit and vegetables influence a goods turnover of shops. For September, 2019 we are faced by a problem of thin setup of a system of video analytics under specifics of shops of different formats. In addition to a goods turnover, system implementation allows to reveal chronic diseases of separate shop, to work in a new way with refill of the range and also to increase efficiency of shop in peak loads, speaks Dobryakov Sergey, the director of the department of the innovative solutions of directorate of the Big Data X5 Retail Group