diff --git a/libraries/AP_Camera/AP_Camera.cpp b/libraries/AP_Camera/AP_Camera.cpp
index fc5f1063fbb..24d8c5beafe 100644
--- a/libraries/AP_Camera/AP_Camera.cpp
+++ b/libraries/AP_Camera/AP_Camera.cpp
@@ -246,6 +246,21 @@ MAV_RESULT AP_Camera::handle_command_long(const mavlink_command_long_t &packet)
         }
         take_picture();
         return MAV_RESULT_ACCEPTED;
+    case MAV_CMD_CAMERA_TRACK_POINT:
+        if (set_tracking(TrackingType::POINT, Vector2f{packet.param1, packet.param2}, Vector2f{})) {
+            return MAV_RESULT_ACCEPTED;
+        }
+        return MAV_RESULT_UNSUPPORTED;
+    case MAV_CMD_CAMERA_TRACK_RECTANGLE:
+        if (set_tracking(TrackingType::RECTANGLE, Vector2f{packet.param1, packet.param2}, Vector2f{packet.param3, packet.param4})) {
+            return MAV_RESULT_ACCEPTED;
+        }
+        return MAV_RESULT_UNSUPPORTED;
+    case MAV_CMD_CAMERA_STOP_TRACKING:
+        if (set_tracking(TrackingType::NONE, Vector2f{}, Vector2f{})) {
+            return MAV_RESULT_ACCEPTED;
+        }
+        return MAV_RESULT_UNSUPPORTED;
     case MAV_CMD_VIDEO_START_CAPTURE:
     case MAV_CMD_VIDEO_STOP_CAPTURE:
     {
@@ -458,6 +473,35 @@ bool AP_Camera::set_focus(uint8_t instance, FocusType focus_type, float focus_va
     return backend->set_focus(focus_type, focus_value);
 }
 
+// set tracking to none, point or rectangle (see TrackingType enum)
+// if POINT only p1 is used, if RECTANGLE then p1 is top-left, p2 is bottom-right
+// p1,p2 are in range 0 to 1.  0 is left or top, 1 is right or bottom
+bool AP_Camera::set_tracking(TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2)
+{
+    WITH_SEMAPHORE(_rsem);
+
+    if (primary == nullptr) {
+        return false;
+    }
+    return primary->set_tracking(tracking_type, p1, p2);
+}
+
+// set tracking to none, point or rectangle (see TrackingType enum)
+// if POINT only p1 is used, if RECTANGLE then p1 is top-left, p2 is bottom-right
+// p1,p2 are in range 0 to 1.  0 is left or top, 1 is right or bottom
+bool AP_Camera::set_tracking(uint8_t instance, TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2)
+{
+    WITH_SEMAPHORE(_rsem);
+
+    auto *backend = get_instance(instance);
+    if (backend == nullptr) {
+        return false;
+    }
+
+    // call each instance
+    return backend->set_tracking(tracking_type, p1, p2);
+}
+
 #if AP_CAMERA_SCRIPTING_ENABLED
 // accessor to allow scripting backend to retrieve state
 // returns true on success and cam_state is filled in
diff --git a/libraries/AP_Camera/AP_Camera.h b/libraries/AP_Camera/AP_Camera.h
index 99d28b91293..78da3ffedee 100644
--- a/libraries/AP_Camera/AP_Camera.h
+++ b/libraries/AP_Camera/AP_Camera.h
@@ -123,6 +123,12 @@ public:
     bool set_focus(FocusType focus_type, float focus_value);
     bool set_focus(uint8_t instance, FocusType focus_type, float focus_value);
 
+    // set tracking to none, point or rectangle (see TrackingType enum)
+    // if POINT only p1 is used, if RECTANGLE then p1 is top-left, p2 is bottom-right
+    // p1,p2 are in range 0 to 1.  0 is left or top, 1 is right or bottom
+    bool set_tracking(TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2);
+    bool set_tracking(uint8_t instance, TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2);
+
     // set if vehicle is in AUTO mode
     void set_is_auto_mode(bool enable) { _is_in_auto_mode = enable; }
 
@@ -135,6 +141,12 @@ public:
         float zoom_value;       // percentage or zoom out = -1, hold = 0, zoom in = 1
         uint8_t focus_type;     // see FocusType enum (1:Rate, 2:Pct, 4:Auto)
         float focus_value;      // If Rate, focus in = -1, focus hold = 0, focus out = 1.  If PCT 0 to 100
+        uint8_t tracking_type;  // see TrackingType enum (0:NONE, 1:POINT, 2:RECTANGLE)
+        float tracking_p1x;     // center or top-left tracking point's x-axis value (0 is left, 1 is right)
+        float tracking_p1y;     // center or top-left tracking point's y-axis value (0 is top, 1 is bottom)
+        float tracking_p2x;     // bottom-right tracking point's x-axis value (0 is left, 1 is right)
+        float tracking_p2y;     // bottom-right tracking point's y-axis value (0 is top, 1 is bottom)
+        Vector2f tracking_p2;   // bottom-right tracking point (0 is left or top, 1 is right or bottom)
     } camera_state_t;
 
     // accessor to allow scripting backend to retrieve state
diff --git a/libraries/AP_Camera/AP_Camera_Backend.h b/libraries/AP_Camera/AP_Camera_Backend.h
index 32940d9e345..2551b9909c2 100644
--- a/libraries/AP_Camera/AP_Camera_Backend.h
+++ b/libraries/AP_Camera/AP_Camera_Backend.h
@@ -65,6 +65,11 @@ public:
     // focus in = -1, focus hold = 0, focus out = 1
     virtual bool set_focus(FocusType focus_type, float focus_value) { return false; }
 
+    // set tracking to none, point or rectangle (see TrackingType enum)
+    // if POINT only p1 is used, if RECTANGLE then p1 is top-left, p2 is bottom-right
+    // p1,p2 are in range 0 to 1.  0 is left or top, 1 is right or bottom
+    virtual bool set_tracking(TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2) { return false; }
+
     // handle MAVLink messages from the camera
     virtual void handle_message(mavlink_channel_t chan, const mavlink_message_t &msg) {}
 
diff --git a/libraries/AP_Camera/AP_Camera_Scripting.cpp b/libraries/AP_Camera/AP_Camera_Scripting.cpp
index f3afb009483..309b712f45c 100644
--- a/libraries/AP_Camera/AP_Camera_Scripting.cpp
+++ b/libraries/AP_Camera/AP_Camera_Scripting.cpp
@@ -37,6 +37,19 @@ bool AP_Camera_Scripting::set_focus(FocusType focus_type, float focus_value)
     return true;
 }
 
+// set tracking to none, point or rectangle (see TrackingType enum)
+// if POINT only p1 is used, if RECTANGLE then p1 is top-left, p2 is bottom-right
+// p1,p2 are in range 0 to 1.  0 is left or top, 1 is right or bottom
+bool AP_Camera_Scripting::set_tracking(TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2)
+{
+    _cam_state.tracking_type = (uint8_t)tracking_type;
+    _cam_state.tracking_p1x = p1.x;
+    _cam_state.tracking_p1y = p1.y;
+    _cam_state.tracking_p2x = p2.x;
+    _cam_state.tracking_p2y = p2.y;
+    return true;
+}
+
 // access for scripting backend to retrieve state
 // returns true on success and cam_state is filled in
 bool AP_Camera_Scripting::get_state(AP_Camera::camera_state_t& cam_state)
diff --git a/libraries/AP_Camera/AP_Camera_Scripting.h b/libraries/AP_Camera/AP_Camera_Scripting.h
index 4e056019715..5cd19a5137a 100644
--- a/libraries/AP_Camera/AP_Camera_Scripting.h
+++ b/libraries/AP_Camera/AP_Camera_Scripting.h
@@ -46,6 +46,11 @@ public:
     // focus in = -1, focus hold = 0, focus out = 1
     bool set_focus(FocusType focus_type, float focus_value) override;
 
+    // set tracking to none, point or rectangle (see TrackingType enum)
+    // if POINT only p1 is used, if RECTANGLE then p1 is top-left, p2 is bottom-right
+    // p1,p2 are in range 0 to 1.  0 is left or top, 1 is right or bottom
+    bool set_tracking(TrackingType tracking_type, const Vector2f& p1, const Vector2f& p2) override;
+
     // returns true on success and cam_state is filled in
     bool get_state(AP_Camera::camera_state_t& cam_state) override;
 
diff --git a/libraries/AP_Camera/AP_Camera_shareddefs.h b/libraries/AP_Camera/AP_Camera_shareddefs.h
index 1687c71228f..dbf473450a8 100644
--- a/libraries/AP_Camera/AP_Camera_shareddefs.h
+++ b/libraries/AP_Camera/AP_Camera_shareddefs.h
@@ -19,3 +19,10 @@ enum class FocusType : uint8_t {
     PCT = 2,    // focus to a percentage (from 0 to 100) of the full range. Same as FOCUS_TYPE_RANGE
     AUTO = 4    // focus automatically. Same as FOCUS_TYPE_AUTO
 };
+
+// tracking types when tracking an object in the video stream
+enum class TrackingType : uint8_t {
+    NONE = 0,       // tracking is inactive
+    POINT = 1,      // tracking a point
+    RECTANGLE = 2   // tracking a rectangle
+};