This study investigates the relationship between pedestrian behavior and the spatial usage of public space in Venice, based on the urban research methodology. By applying computer vision techniques—including object detection and trajectory reconstruction—to time-series imagery captured from fixed observation points, the study visualizes human activity in a public campo. The extracted behavior data is spatially mapped to generate dynamic behavioral maps that reflect how everyday use patterns shape and reveal the quality of public space. This study enables the application of visual data in geospatial analysis by distinguishing types of behavior in short-term image sequences, providing a new method for quantifying the use of public spaces. By integrating urban morphology, urban context, behavior type, and time pattern, the study enables a layered spatial interpretation of public environments. In discussions about people-centered spaces, how to objectively, continuously, and visually understand crowd behavior has become an important issue. Based on traditional cognitive behavior maps, this study proposes a method of reconstructing the cognition of public spaces through a computational method. Although computer vision technology has previously been used to visualize urban dynamics and detect people, this study is the first to distinguish pedestrian behavior through short-term continuous image capture, thereby quantifying the use of public spaces in Venice.