A heuristic approach to evaluate occurrences of products for the planogram maintenance

Planogram is a detailed visual map of products in a retail store and establishes the position of products in order to increase sales and to supply the best location for suppliers. So, the aims of “correct” planogram are several and are: increasing sales, increasing profits, introducing a new item, supporting an innovative merchandising approach, and better manages the shelves. Deviating from the planogram defeats the purpose of any of those goals. A fundamental aspect in the retail operations is to maintain the integrity of the planogram. This work intends to provide a solution to this problem, proposing a system that individually identifies the presence of a specific product in the image of a shelf. This even though the product is moved, rotated, misplaced, or even in poor lighting conditions. This paper presents a method to find and count multiple instances of the same object that occurs into an image of a shelf in a store without using classifiers. The procedures here described are based on a heuristic algorithm that involves morphological operation, template matching and histogram comparison. Experimental results are presented in order to verify the effectiveness of the proposed approach. They demonstrate that the algorithm provides satisfactory results when the user manually chooses the most significant label of the product to find in the shelf image.