This paper presents a method for modeling the interaction of light between diffusely reflecting surfaces. Current light reflection models in computer graphics do not account for object-to-object reflections between diffuse surfaces, leading to incorrect global illumination effects. The proposed method, based on thermal engineering principles, includes the effects of diffuse light sources of finite area and "color-bleeding" caused by diffuse reflections. It is applicable to environments composed of ideal diffuse reflectors and can account for direct illumination from various light sources. The resultant surface intensities are independent of observer position, allowing preprocessing for dynamic sequences. The method is based on energy principles and can be applied monochromatically or to finite wavelength intervals. The key assumption is that all surfaces are ideal diffuse (Lambertian) reflectors. The procedure accounts for direct illumination and multiple reflections within the environment. A major advantage is that surface intensities are independent of observer position, enabling preprocessing for dynamic sequences. The method involves calculating the radiosity of each surface, which is the hemispherical integral of the energy leaving the surface. Radiosity is expressed as the sum of direct emission and reflected light. The radiosity method accounts for all light leaving and incident upon a surface, and the incident light is expressed in terms of incident radiosity. The method is applicable to any environment with diffuse surfaces and can handle multiple reflections. The method was implemented in a program that reads an environment description, subdivides polygons into subpolygon elements, computes form factors between elements, and solves the matrix version of the radiosity equation to obtain element intensities. The program is limited to convex, polygonal surfaces and does not account for hidden surfaces. The results show that the radiosity method accurately simulates the interaction of light between diffuse surfaces, including color-bleeding effects. The method was compared with a physical model, and the results showed good agreement. The method is computationally expensive but has the advantage of being independent of observer position, allowing preprocessing for dynamic sequences. Future work includes creating a smarter subdivision algorithm and considering occluded surfaces and non-polygonal objects.This paper presents a method for modeling the interaction of light between diffusely reflecting surfaces. Current light reflection models in computer graphics do not account for object-to-object reflections between diffuse surfaces, leading to incorrect global illumination effects. The proposed method, based on thermal engineering principles, includes the effects of diffuse light sources of finite area and "color-bleeding" caused by diffuse reflections. It is applicable to environments composed of ideal diffuse reflectors and can account for direct illumination from various light sources. The resultant surface intensities are independent of observer position, allowing preprocessing for dynamic sequences. The method is based on energy principles and can be applied monochromatically or to finite wavelength intervals. The key assumption is that all surfaces are ideal diffuse (Lambertian) reflectors. The procedure accounts for direct illumination and multiple reflections within the environment. A major advantage is that surface intensities are independent of observer position, enabling preprocessing for dynamic sequences. The method involves calculating the radiosity of each surface, which is the hemispherical integral of the energy leaving the surface. Radiosity is expressed as the sum of direct emission and reflected light. The radiosity method accounts for all light leaving and incident upon a surface, and the incident light is expressed in terms of incident radiosity. The method is applicable to any environment with diffuse surfaces and can handle multiple reflections. The method was implemented in a program that reads an environment description, subdivides polygons into subpolygon elements, computes form factors between elements, and solves the matrix version of the radiosity equation to obtain element intensities. The program is limited to convex, polygonal surfaces and does not account for hidden surfaces. The results show that the radiosity method accurately simulates the interaction of light between diffuse surfaces, including color-bleeding effects. The method was compared with a physical model, and the results showed good agreement. The method is computationally expensive but has the advantage of being independent of observer position, allowing preprocessing for dynamic sequences. Future work includes creating a smarter subdivision algorithm and considering occluded surfaces and non-polygonal objects.