Photogrammery

[OrionOth]

Here lies my notes on making a to-scale 3d model of a specific pvc pipe structure.

Rough OpenSCAD mockup of the shape

WIDTH = 36; // 36 cm
UNION_OFFSET = 0.001;
LABEL_SIZE = 3;
LABEL_FONT = "monospace";
LABEL_COLOR = "black";

x1 = 35; // unknown
y1 = 10 + 4;
x2 = 32 + 4; // corner to corner inside pipes
y2 = 26 + 2 + 30; // unknown
x3 = 40; // unknown
y3 = 26 + 4;

ch = 15; // unknown
rh = 5;  // irrelevant abstraction

//p0 = [0, 0];
p1 = [x1 + x2 + x3, 0];
p2 = [x1 + x2 + x3, y3];
p3 = [x1 + x2, y3];
p4 = [x1 + x2, y2];
p5 = [x1, y2];
p6 = [x1, y1];
p7 = [0, y1];

translate([-(x1 + x2/2), 0, 0])
rotate([90,0,0]){
    linear_extrude(WIDTH, center = true)
        polygon(points = [[0, 0], p1, p2, p3, p4, p5, p6, p7]);

    // Labels
    translate([0,0,WIDTH/2])
    color(LABEL_COLOR)
    linear_extrude(1, center = true)
    {
        translate([x1/2,y1,0])
            text(str(x1, "cm"), LABEL_SIZE, LABEL_FONT, halign = "center", valign = "top");
        translate([0,y1/2,0])
            text(str(y1, "cm"), LABEL_SIZE, LABEL_FONT, halign = "left", valign = "center");
        translate([x1+x2/2, y2, 0])
            text(str(x2, "cm"), LABEL_SIZE, LABEL_FONT, halign = "center", valign = "top");
        translate([x1,y1+(y2-y1)/2,0])
            text(str(y2-y1, "cm"), LABEL_SIZE, LABEL_FONT, halign = "left", valign = "center");
        translate([x1+x2+x3/2,y3,0])
            text(str(x3, "cm"), LABEL_SIZE, LABEL_FONT, halign = "center", valign = "top");
        translate([x1+x2,y3+(y2-y3)/2,0])
            text(str(y2-y3, "cm"), LABEL_SIZE, LABEL_FONT, halign = "right", valign = "center");
        translate([x1+x2+x3,y3/2,0])
            text(str(y3, "cm"), LABEL_SIZE, LABEL_FONT, halign = "right", valign = "center");
        translate([(x1+x2+x3)/2,0,0])
            text(str(x1+x2+x3, "cm"), LABEL_SIZE, LABEL_FONT, halign = "center", valign = "bottom");
    }
}

// Bounding box for top coral
translate([0,0,y2 + ch/2])
cylinder(h = ch, r = rh, center = true);

Keywords: Photogrammetry, Structure from Motion (SFM), Multi-view Geometry (MVG), Computer Vision (CV), Simultaneous Localization and Mapping (SLAM), Scale-invariant feature transform (SIFT)

https://szeliski.org/Book/