In an IEEE Very Small Size Soccer match, an image capture system uses a camera positioned above the pitch to monitor the location of the game elements, such as the robots and the ball. These images are processed by a computer, which sends commands to the robots, enabling the game to proceed autonomously. This article presents the initial development of a computer vision system for mobile robots in this category. The work's main focus is capturing, processing, and recognizing data from the images captured by the camera. This information was processed using algorithms in the Python programming language, with the help of the OpenCV computer vision library. This tool contains the HSV color system to work with the parameters of tone, saturation, illumination, and brightness of the images. Based on these features, it was possible to develop image-processing methods for object identification, obtaining promising results for the continuation of the project.