PinSolver：Blender 批量相机与图像序列插件

Overview

PinSolver is a Blender add-on that leverages OpenCV's powerful PnP (Perspective-n-Point) algorithm to provide intuitive, fast camera and object alignment. By linking 2D coordinates (●) on an image to 3D coordinates (■) in 3D space, it seamlessly supports everything from static object placement to full-fledged video matchmoving.

Prerequisites

• Accurately Scaled Models: To achieve high-precision alignment and matchmoving, it is absolutely critical that the 3D models or scan data placed in your scene match real-world proportions and dimensions. Incorrect model scale will yield inaccurate calculation results.

Use Cases

• VFX Compositing, Layout, and Survey Work for Live-Action: Useful for camera layouts against background models, aligning with scan data, and placing props. It also offers a streamlined matchmoving feature using PnP to perform camera solves based on Blender's tracking data.
• Camera projection. Texture projection.
• Aligning architectural renderings
• Aligning models with reference photos
• Lens Parameter Estimation: Even if the camera metadata is unknown, PinSolver automatically estimates the focal length, optical center, and lens distortion based on the spatial relationship between the pins and the 3D model. Matchmove mode also supports variable zoom. (Note: Variable animated distortion is not supported due to Blender's internal constraints.)

Features

🔀 Dual Workflow Modes

PinSolver includes two distinct modes depending on your objective:

• Layout Mode: Dedicated to single-frame static camera matching and prop placement.
• Matchmove Mode: Integrates with Blender's 2D tracking data to dynamically bake camera or object animations.

🎯 Core PnP Solver

• 1-Pin (Raycast): Instantly snaps the target along the camera's ray (line of sight) using a single point.
• 2-Pin (Pan/Orbit): Calculates basic orientation and translation using two points.

查看官方完整介绍 ↗