Save the 2D bounding box of an object in rendered image to a text file

Question

I have a blend file with a moving object and want to save the object's bounding box positions (x, y, width and height) in the rendered image to a text file (for every frame):

The bounding box values need to be in the pixelspace of the rendered 2D image, where x and y are the location of the image's pixel representing the top left corner, while widthand height represent the number of pixels from the top left corner to the right bottom corner.

Answer 1

This script determines the camera space bounding box and calculates the top left corner location in the rendered image + width and height (clamped to render resolution).

It will return incorrect results if the object is partially behind the camera, it should work if it's in front and crossing the camera view border however.

import bpy
from mathutils import Vector

class Box:

    dim_x = 1
    dim_y = 1

    def __init__(self, min_x, min_y, max_x, max_y, dim_x=dim_x, dim_y=dim_y):
        self.min_x = min_x
        self.min_y = min_y
        self.max_x = max_x
        self.max_y = max_y
        self.dim_x = dim_x
        self.dim_y = dim_y

    @property
    def x(self):
        return round(self.min_x * self.dim_x)

    @property
    def y(self):
        return round(self.dim_y - self.max_y * self.dim_y)

    @property
    def width(self):
        return round((self.max_x - self.min_x) * self.dim_x)

    @property
    def height(self):
        return round((self.max_y - self.min_y) * self.dim_y)

    def __str__(self):
        return "<Box, x=%i, y=%i, width=%i, height=%i>" % \
               (self.x, self.y, self.width, self.height)

    def to_tuple(self):
        if self.width == 0 or self.height == 0:
            return (0, 0, 0, 0)
        return (self.x, self.y, self.width, self.height)


def camera_view_bounds_2d(scene, cam_ob, me_ob):
    """
    Returns camera space bounding box of mesh object.

    Negative 'z' value means the point is behind the camera.

    Takes shift-x/y, lens angle and sensor size into account
    as well as perspective/ortho projections.

    :arg scene: Scene to use for frame size.
    :type scene: :class:`bpy.types.Scene`
    :arg obj: Camera object.
    :type obj: :class:`bpy.types.Object`
    :arg me: Untransformed Mesh.
    :type me: :class:`bpy.types.Mesh´
    :return: a Box object (call its to_tuple() method to get x, y, width and height)
    :rtype: :class:`Box`
    """

    mat = cam_ob.matrix_world.normalized().inverted()
    me = me_ob.to_mesh(scene, True, 'PREVIEW')
    me.transform(me_ob.matrix_world)
    me.transform(mat)

    camera = cam_ob.data
    frame = [-v for v in camera.view_frame(scene=scene)[:3]]
    camera_persp = camera.type != 'ORTHO'

    lx = []
    ly = []

    for v in me.vertices:
        co_local = v.co
        z = -co_local.z

        if camera_persp:
            if z == 0.0:
                lx.append(0.5)
                ly.append(0.5)
            # Does it make any sense to drop these?
            #if z <= 0.0:
            #    continue
            else:
                frame = [(v / (v.z / z)) for v in frame]

        min_x, max_x = frame[1].x, frame[2].x
        min_y, max_y = frame[0].y, frame[1].y

        x = (co_local.x - min_x) / (max_x - min_x)
        y = (co_local.y - min_y) / (max_y - min_y)

        lx.append(x)
        ly.append(y)

    min_x = clamp(min(lx), 0.0, 1.0)
    max_x = clamp(max(lx), 0.0, 1.0)
    min_y = clamp(min(ly), 0.0, 1.0)
    max_y = clamp(max(ly), 0.0, 1.0)

    bpy.data.meshes.remove(me)

    r = scene.render
    fac = r.resolution_percentage * 0.01
    dim_x = r.resolution_x * fac
    dim_y = r.resolution_y * fac

    return Box(min_x, min_y, max_x, max_y, dim_x, dim_y)


def clamp(x, minimum, maximum):
    return max(minimum, min(x, maximum))


def write_bounds_2d(filepath, scene, cam_ob, me_ob, frame_start, frame_end):

    with open(filepath, "w") as file: 
        for frame in range(frame_start, frame_end + 1):
            bpy.context.scene.frame_set(frame)
            file.write("%i %i %i %i\n" % camera_view_bounds_2d(scene, cam_ob, me_ob).to_tuple())

def main(context):

    filepath = r"D:\temp\bounds_2d.txt"

    scene = context.scene
    cam_ob = scene.camera
    me_ob = context.object

    frame_current = scene.frame_current
    frame_start = scene.frame_start
    frame_end = scene.frame_end

    write_bounds_2d(filepath, scene, cam_ob, me_ob, frame_start, frame_end)

    scene.frame_set(frame_current)

main(bpy.context)

One could try cut the mesh at the camera view border (world space, camera space or NDC?) and throw away all coordinates outside the frame. That should fix the bounding box problem with object partially behind camera. If we dropped these coords without cutting, it would likely return a too small box (should return a too large box in these cases at the moment).

Answer 2

For anyone looking for a Blender 2.8 version of @CoDEmanX's answer

def clamp(x, minimum, maximum):
    return max(minimum, min(x, maximum))

def camera_view_bounds_2d(scene, cam_ob, me_ob):
    """
    Returns camera space bounding box of mesh object.

    Negative 'z' value means the point is behind the camera.

    Takes shift-x/y, lens angle and sensor size into account
    as well as perspective/ortho projections.

    :arg scene: Scene to use for frame size.
    :type scene: :class:`bpy.types.Scene`
    :arg obj: Camera object.
    :type obj: :class:`bpy.types.Object`
    :arg me: Untransformed Mesh.
    :type me: :class:`bpy.types.Mesh´
    :return: a Box object (call its to_tuple() method to get x, y, width and height)
    :rtype: :class:`Box`
    """

    mat = cam_ob.matrix_world.normalized().inverted()
    depsgraph = bpy.context.evaluated_depsgraph_get()
    mesh_eval = me_ob.evaluated_get(depsgraph)
    me = mesh_eval.to_mesh()
    me.transform(me_ob.matrix_world)
    me.transform(mat)

    camera = cam_ob.data
    frame = [-v for v in camera.view_frame(scene=scene)[:3]]
    camera_persp = camera.type != 'ORTHO'

    lx = []
    ly = []

    for v in me.vertices:
        co_local = v.co
        z = -co_local.z

        if camera_persp:
            if z == 0.0:
                lx.append(0.5)
                ly.append(0.5)
            # Does it make any sense to drop these?
            # if z <= 0.0:
            #    continue
            else:
                frame = [(v / (v.z / z)) for v in frame]

        min_x, max_x = frame[1].x, frame[2].x
        min_y, max_y = frame[0].y, frame[1].y

        x = (co_local.x - min_x) / (max_x - min_x)
        y = (co_local.y - min_y) / (max_y - min_y)

        lx.append(x)
        ly.append(y)

    min_x = clamp(min(lx), 0.0, 1.0)
    max_x = clamp(max(lx), 0.0, 1.0)
    min_y = clamp(min(ly), 0.0, 1.0)
    max_y = clamp(max(ly), 0.0, 1.0)

    mesh_eval.to_mesh_clear()

    r = scene.render
    fac = r.resolution_percentage * 0.01
    dim_x = r.resolution_x * fac
    dim_y = r.resolution_y * fac

    # Sanity check
    if round((max_x - min_x) * dim_x) == 0 or round((max_y - min_y) * dim_y) == 0:
        return (0, 0, 0, 0)

    return (
        round(min_x * dim_x),            # X
        round(dim_y - max_y * dim_y),    # Y
        round((max_x - min_x) * dim_x),  # Width
        round((max_y - min_y) * dim_y)   # Height
    )

# Print the result
print(camera_view_bounds_2d(context.scene, context.scene.camera, context.object))