Labels are text annotations of graphical objects in visualization diagrams. Label placement methods automatically find appropriate positions for graphical objects to place text annotations while achieving high readability and unambiguity of labels. The methods have evolved over the past half century. In recent years, the wide application of data-driven visualization generation technology has promoted the rapid development of label placement methods. We survey relevant literatures of the last 50 years. Taking the three types of graphical objects, namely point, line, and area, as the main perspective, we summarize existing label placement methods by graphical object type from three aspects: label candidate location models, label placement algorithms, and label placement evaluation indicators. Furthermore, we discuss on the prospects of challenges and directions for the future research of label placement methods.