Parcel volume measurement is a crucial aspect of logistics billing. Addressing the issues of low real-time performance and unreliable accuracy under parcel stacking scenarios in existing volume measurement systems, this paper proposes a real-time parcel volume measurement method based on depth map collision line tracing. The method comprises three main components: collision line tracing, conditional depth map completion, and steep surface grid volume calibration. Initially, RGB and depth map data of parcels are captured in real time through a binocular camera from a top-down perspective. The collision line tracing algorithm is employed to optimize the repeated measurement issue in parcel stacking conditions. Subsequently, a conditional depth map completion strategy is utilized to address the loss of depth data due to factors such as material properties and hardware limitations. Finally, the steep surface grid volume calibration technique is applied to enhance the measurement precision of the grid volume at the parcel edges. The experimental results demonstrate that the algorithm achieves a volume measurement accuracy of over 98% for both regular and irregular parcels under three different speed conditions: 0.5 m/s, 1 m/s, and 2 m/s. The average deviation is less than 2%, indicating high measurement precision and robustness. The algorithm fully meets the logistics industry’s requirements for real-time, high-precision measurements.