Material gathering activities are among the most time consuming in warehouse management. Work-in-Progress (WIP) inventory is stored in the factory floor in between manufacturing operations and often obstructs efficient material gathering by barricading efficient routes from one point to another. Additionally, the WIP inventory is can be placed in different locations and the placement can change dynamically. This research proposes to combine traditional RFID technologies and computer vision to identify most up-to-date locations of the WIP inventory and to use this information for guiding material pick-up activities. To achieve that, a layout of manufacturing facilities is captured and dynamically updated using sensing technologies. These data are stored in a spatial database, and routing algorithms use the spatial data to find optimal routes in crowded manufacturing facilities. The paper describes the general approach to spatial modeling and the proposed database schema and illustrates application of the approach at real manufacturing facilities.