研究目的
To estimate the positions and intensities of multiple near point light sources from a single view image using feature points in cast shadows and specular highlights.
研究成果
The proposed method for multiple near point light sources estimation uses information from specular highlights, feature points in cast shadows and diffuse component on the Lambertian ground plane. Experimental results have been reported on the method’s effectiveness.
研究不足
The method assumes that the camera parameter and the geometry of objects in the scene are known, the scene is illuminated by one or more point light sources, all specular peaks corresponding to light sources are visible in the image used for the estimation, all cast shadows on the floor due all light sources are visible on the taken image, the reflectance properties of the spherical and plane objects are known, the floor plane has a Lambertian surface, and objects in the scene contain one or more surface discontinuity point.
1：Experimental Design and Method Selection:
The method utilizes a specular highlight from an object of known geometry and location (a sphere in this case), and a ground plane with diffusive reflection. The spherical object is required to contain some discontinuity on its surface.
2：Sample Selection and Data Sources:
Synthetic images at a resolution of 1000×1000 pixels, created using Blender 3D software.
3：List of Experimental Equipment and Materials:
A camera, a spherical object with a small conic shape on top, and up to three point light sources.
4：Experimental Procedures and Operational Workflow:
The method involves finding specular peak pixels, calculating the directions of light sources, detecting possible corner points on the shadow, pairing each candidate corner point with the corresponding discontinuity point on the object, and choosing light source position candidates that best fit the ground-plane diffused light equation.
5：Data Analysis Methods:
The Root Means Square Error (RMSE) of the estimated light position is calculated and used as accuracy measure.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
