Add output attribute to camera elements.

This attribute allows specifying the intended output types (e.g., RGB, depth, normal) for each camera in the XML. The output types are stored as bit flags in `mjModel`.

PiperOrigin-RevId: 853316261
Change-Id: I1ad2e94102593cf402756f8f4c80509371ca86d2

src/user/user_init.c

@@ -21,7 +21,6 @@
 #include "user/user_api.h"
 
 
-
 // default model attributes
 void mjs_defaultSpec(mjSpec* spec) {
   memset(spec, 0, sizeof(mjSpec));
@@ -66,14 +65,12 @@ void mjs_defaultSpec(mjSpec* spec) {
 }
 
 
-
 // default orientation attributes
 void mjs_defaultOrientation(mjsOrientation* orient) {
   memset(orient, 0, sizeof(mjsOrientation));
 }
 
 
-
 // default body attributes
 void mjs_defaultBody(mjsBody* body) {
   memset(body, 0, sizeof(mjsBody));
@@ -88,7 +85,6 @@ void mjs_defaultBody(mjsBody* body) {
 }
 
 
-
 // default frame attributes
 void mjs_defaultFrame(mjsFrame* frame) {
   memset(frame, 0, sizeof(mjsFrame));
@@ -96,7 +92,6 @@ void mjs_defaultFrame(mjsFrame* frame) {
 }
 
 
-
 // default joint attributes
 void mjs_defaultJoint(mjsJoint* joint) {
   memset(joint, 0, sizeof(mjsJoint));
@@ -110,7 +105,6 @@ void mjs_defaultJoint(mjsJoint* joint) {
 }
 
 
-
 // default geom attributes
 void mjs_defaultGeom(mjsGeom* geom) {
   memset(geom, 0, sizeof(mjsGeom));
@@ -153,7 +147,6 @@ void mjs_defaultGeom(mjsGeom* geom) {
 }
 
 
-
 // default site attributes
 void mjs_defaultSite(mjsSite* site) {
   memset(site, 0, sizeof(mjsSite));
@@ -172,7 +165,6 @@ void mjs_defaultSite(mjsSite* site) {
 }
 
 
-
 // default cam attributes
 void mjs_defaultCamera(mjsCamera* cam) {
   memset(cam, 0, sizeof(mjsCamera));
@@ -187,10 +179,10 @@ void mjs_defaultCamera(mjsCamera* cam) {
   cam->fovy = 45;
   cam->ipd = 0.068;
   cam->resolution[0] = cam->resolution[1] = 1;
+  cam->output = mjCAMOUT_RGB;
 }
 
 
-
 // default light attributes
 void mjs_defaultLight(mjsLight* light) {
   memset(light, 0, sizeof(mjsLight));
@@ -215,7 +207,6 @@ void mjs_defaultLight(mjsLight* light) {
 }
 
 
-
 // default flex attributes
 void mjs_defaultFlex(mjsFlex* flex) {
   memset(flex, 0, sizeof(mjsFlex));
@@ -242,7 +233,6 @@ void mjs_defaultFlex(mjsFlex* flex) {
 }
 
 
-
 // default mesh attributes
 void mjs_defaultMesh(mjsMesh* mesh) {
   memset(mesh, 0, sizeof(mjsMesh));
@@ -253,14 +243,12 @@ void mjs_defaultMesh(mjsMesh* mesh) {
 }
 
 
-
 // default height field attributes
 void mjs_defaultHField(mjsHField* hfield) {
   memset(hfield, 0, sizeof(mjsHField));
 }
 
 
-
 // default skin attributes
 void mjs_defaultSkin(mjsSkin* skin) {
   memset(skin, 0, sizeof(mjsSkin));
@@ -269,7 +257,6 @@ void mjs_defaultSkin(mjsSkin* skin) {
 }
 
 
-
 // default texture attributes
 void mjs_defaultTexture(mjsTexture* texture) {
   memset(texture, 0, sizeof(mjsTexture));
@@ -285,7 +272,6 @@ void mjs_defaultTexture(mjsTexture* texture) {
 }
 
 
-
 // default material attributes
 void mjs_defaultMaterial(mjsMaterial* material) {
   memset(material, 0, sizeof(mjsMaterial));
@@ -298,7 +284,6 @@ void mjs_defaultMaterial(mjsMaterial* material) {
 }
 
 
-
 // default pair attributes
 void mjs_defaultPair(mjsPair* pair) {
   memset(pair, 0, sizeof(mjsPair));
@@ -312,7 +297,6 @@ void mjs_defaultPair(mjsPair* pair) {
 }
 
 
-
 // default equality attributes
 void mjs_defaultEquality(mjsEquality* equality) {
   memset(equality, 0, sizeof(mjsEquality));
@@ -324,7 +308,6 @@ void mjs_defaultEquality(mjsEquality* equality) {
 }
 
 
-
 // default tendon attributes
 void mjs_defaultTendon(mjsTendon* tendon) {
   memset(tendon, 0, sizeof(mjsTendon));
@@ -338,7 +321,6 @@ void mjs_defaultTendon(mjsTendon* tendon) {
 }
 
 
-
 // default actuator attributes
 void mjs_defaultActuator(mjsActuator* actuator) {
   memset(actuator, 0, sizeof(mjsActuator));
@@ -364,7 +346,6 @@ void mjs_defaultActuator(mjsActuator* actuator) {
 }
 
 
-
 // default sensor attributes
 void mjs_defaultSensor(mjsSensor* sensor) {
   memset(sensor, 0, sizeof(mjsSensor));
@@ -375,35 +356,30 @@ void mjs_defaultSensor(mjsSensor* sensor) {
 }
 
 
-
 // Default numeric attributes.
 void mjs_defaultNumeric(mjsNumeric* numeric) {
   memset(numeric, 0, sizeof(mjsNumeric));
 }
 
 
-
 // Default text attributes.
 void mjs_defaultText(mjsText* text) {
   memset(text, 0, sizeof(mjsText));
 }
 
 
-
 // Default tuple attributes.
 void mjs_defaultTuple(mjsTuple* tuple) {
   memset(tuple, 0, sizeof(mjsTuple));
 }
 
 
-
 // Default keyframe attributes.
 void mjs_defaultKey(mjsKey* key) {
   memset(key, 0, sizeof(mjsKey));
 }
 
 
-
 // default plugin attributes
 void mjs_defaultPlugin(mjsPlugin* plugin) {
   memset(plugin, 0, sizeof(mjsPlugin));

src/user/user_model.cc

@@ -2868,6 +2868,7 @@ void mjCModel::CopyTree(mjModel* m) {
       m->cam_fovy[cid] = (mjtNum)pc->fovy;
       m->cam_ipd[cid] = (mjtNum)pc->ipd;
       mjuu_copyvec(m->cam_resolution+2*cid, pc->resolution, 2);
+      m->cam_output[cid] = pc->output;
       mjuu_copyvec(m->cam_sensorsize+2*cid, pc->sensor_size, 2);
       mjuu_copyvec(m->cam_intrinsic+4*cid, pc->intrinsic, 4);
       mjuu_copyvec(m->cam_user+nuser_cam*cid, pc->get_userdata().data(), nuser_cam);

src/xml/xml_base.h

@@ -45,6 +45,7 @@ extern const int gain_sz;
 extern const int bias_sz;
 extern const int stage_sz;
 extern const int datatype_sz;
+extern const int camout_sz;
 extern const int reduce_sz;
 extern const mjMap angle_map[];
 extern const mjMap enable_map[];
@@ -77,6 +78,7 @@ extern const mjMap stage_map[];
 extern const mjMap datatype_map[];
 extern const mjMap condata_map[];
 extern const mjMap raydata_map[];
+extern const mjMap camout_map[];
 extern const mjMap reduce_map[];
 extern const mjMap meshtype_map[];
 extern const mjMap meshinertia_map[];

src/xml/xml_native_reader.cc

@@ -171,7 +171,7 @@ std::vector<const char*> MJCF[nMJCF] = {
             "hfield", "mesh", "fitscale", "rgba", "fluidshape", "fluidcoef", "user"},
         {"site", "?", "type", "group", "pos", "quat", "material",
             "size", "fromto", "axisangle", "xyaxes", "zaxis", "euler", "rgba", "user"},
-        {"camera", "?", "projection", "fovy", "ipd", "resolution", "pos", "quat",
+        {"camera", "?", "projection", "fovy", "ipd", "resolution", "output", "pos", "quat",
             "axisangle", "xyaxes", "zaxis", "euler", "mode", "focal", "focalpixel",
             "principal", "principalpixel", "sensorsize", "user"},
         {"light", "?", "pos", "dir", "bulbradius", "intensity", "range",
@@ -285,7 +285,7 @@ std::vector<const char*> MJCF[nMJCF] = {
         {"attach", "*", "model", "body", "prefix"},
         {"site", "*",  "name", "class", "type", "group", "pos", "quat",
             "material", "size", "fromto", "axisangle", "xyaxes", "zaxis", "euler", "rgba", "user"},
-        {"camera", "*", "name", "class", "projection", "fovy", "ipd", "resolution", "pos",
+        {"camera", "*", "name", "class", "projection", "fovy", "ipd", "resolution", "output", "pos",
             "quat", "axisangle", "xyaxes", "zaxis", "euler", "mode", "target",
             "focal", "focalpixel", "principal", "principalpixel", "sensorsize", "user"},
         {"light", "*", "name", "class", "directional", "type", "castshadow", "active",
@@ -793,6 +793,13 @@ const mjMap raydata_map[mjNRAYDATA] = {
   {"depth",         mjRAYDATA_DEPTH}
 };
 
+// camera output type
+const int camout_sz = mjNCAMOUT;
+const mjMap camout_map[mjNCAMOUT] = {{"rgb", mjCAMOUT_RGB},
+                                     {"depth", mjCAMOUT_DEPTH},
+                                     {"distance", mjCAMOUT_DIST},
+                                     {"normal", mjCAMOUT_NORMAL},
+                                     {"segmentation", mjCAMOUT_SEG}};
 
 // contact reduction type
 const int reduce_sz = 4;
@@ -1953,6 +1960,17 @@ void mjXReader::OneCamera(XMLElement* elem, mjsCamera* camera) {
   ReadAttr(elem, "focalpixel", 2, camera->focal_pixel, text);
   ReadAttr(elem, "focal", 2, camera->focal_length, text);
   ReadAttr(elem, "resolution", 2, camera->resolution, text);
+
+  // read output attribute as space-separated bitflags
+  std::vector<int> outvals(mjNCAMOUT);
+  int nout = MapValues(elem, "output", outvals.data(), camout_map, mjNCAMOUT);
+  if (nout) {
+    camera->output = 0;
+    for (int i = 0; i < nout; ++i) {
+      camera->output |= outvals[i];
+    }
+  }
+
   bool sensorsize = ReadAttr(elem, "sensorsize", 2, camera->sensor_size, text);
   bool fovy = ReadAttr(elem, "fovy", 1, &camera->fovy, text);
   if (fovy && sensorsize) {

src/xml/xml_native_writer.cc

@@ -573,6 +573,19 @@ void mjXWriter::OneCamera(XMLElement* elem, const mjCCamera* camera, mjCDef* def
   WriteAttr(elem, "ipd", 1, &camera->ipd, &def->Camera().ipd);
   WriteAttrKey(elem, "mode", camlight_map, camlight_sz, camera->mode, def->Camera().mode);
   WriteAttr(elem, "resolution", 2, camera->resolution, def->Camera().resolution);
+
+  // write output attribute if different from default
+  if (camera->output != def->Camera().output) {
+    int data[mjNCAMOUT];
+    int ndata = 0;
+    for (int i = 0; i < mjNCAMOUT; i++) {
+      if (camera->output & camout_map[i].value) {
+        data[ndata++] = camout_map[i].value;
+      }
+    }
+    WriteAttrKeys(elem, "output", camout_map, camout_sz, data, ndata, 0);
+  }
+
   WriteAttrKey(elem, "projection", projection_map, projection_sz, camera->proj, def->Camera().proj);
 
   // camera intrinsics if specified