SD-VLM: Spatial Measuring and Understanding with Depth-Encoded Vision-Language Models