aruco_pose nodelet cleanup (#239)

* aruco_pose: Unhardcode contour refinement

Besides, this was basically a no-op anyway, since dynamic parameters
overwrote that anyway.

* aruco_pose: Late-construct objects that use ROS

* aruco_map: Don't create/store node handle

* aruco_pose: Don't assume dist_coeffs size

* aruco_pose: more const == more better

* aruco_pose: Be more obvious about changing variables

* aruco_pose: Fix building for Kinetic

* aruco_pose: Remove global add_definitions

aruco_pose/CMakeLists.txt

@@ -1,8 +1,6 @@
 cmake_minimum_required(VERSION 3.0)
 project(aruco_pose)
 
-add_definitions(-std=c++11 -Wall -g)
-
 ## Compile as C++11, supported in ROS Kinetic and newer
 add_compile_options(-std=c++11)
 
@@ -25,7 +23,7 @@ find_package(catkin REQUIRED COMPONENTS
 )
 
 find_package(OpenCV 3 REQUIRED COMPONENTS core imgproc calib3d)
-if ("${OpenCV_VERSION_MINOR}" LESS "3")
+if ("${OpenCV_VERSION_MINOR}" LESS "9")
   message(STATUS "OpenCV version too low, using vendored ArUco package")
   include(vendor/VendorOpenCV.cmake)
 else()
@@ -229,4 +227,5 @@ if (CATKIN_ENABLE_TESTING)
   add_rostest(test/test_parser_empty_map.test)
   add_rostest(test/test_node_failure.test)
   add_rostest(test/largemap.test)
+  add_rostest(test/crash_opencv.test)
 endif()

aruco_pose/src/aruco_detect.cpp

@@ -58,10 +58,9 @@ using cv::Mat;
 
 class ArucoDetect : public nodelet::Nodelet {
 private:
-	ros::NodeHandle nh_, nh_priv_;
-	tf2_ros::TransformBroadcaster br_;
-	tf2_ros::Buffer tf_buffer_;
-	tf2_ros::TransformListener tf_listener_{tf_buffer_};
+	std::unique_ptr<tf2_ros::TransformBroadcaster> br_;
+	std::unique_ptr<tf2_ros::Buffer> tf_buffer_;
+	std::unique_ptr<tf2_ros::TransformListener> tf_listener_;
 	std::shared_ptr<dynamic_reconfigure::Server<aruco_pose::DetectorConfig>> dyn_srv_;
 	cv::Ptr<cv::aruco::Dictionary> dictionary_;
 	cv::Ptr<cv::aruco::DetectorParameters> parameters_;
@@ -81,30 +80,32 @@ private:
 public:
 	virtual void onInit()
 	{
-		nh_ = getNodeHandle();
-		nh_priv_ = getPrivateNodeHandle();
+		ros::NodeHandle& nh_ = getNodeHandle();
+		ros::NodeHandle& nh_priv_ = getPrivateNodeHandle();
+
+		br_.reset(new tf2_ros::TransformBroadcaster());
+		tf_buffer_.reset(new tf2_ros::Buffer());
+		tf_listener_.reset(new tf2_ros::TransformListener(*tf_buffer_, nh_));
 
 		int dictionary;
-		nh_priv_.param("dictionary", dictionary, 2);
-		nh_priv_.param("estimate_poses", estimate_poses_, true);
-		nh_priv_.param("send_tf", send_tf_, true);
+		dictionary = nh_priv_.param("dictionary", 2);
+		estimate_poses_ = nh_priv_.param("estimate_poses", true);
+		send_tf_ = nh_priv_.param("send_tf", true);
 		if (estimate_poses_ && !nh_priv_.getParam("length", length_)) {
 			NODELET_FATAL("can't estimate marker's poses as ~length parameter is not defined");
 			ros::shutdown();
 		}
-		readLengthOverride();
+		readLengthOverride(nh_priv_);
 
-		nh_priv_.param<std::string>("known_tilt", known_tilt_, "");
-		nh_priv_.param("auto_flip", auto_flip_, false);
+		known_tilt_ = nh_priv_.param<std::string>("known_tilt", "");
+		auto_flip_ = nh_priv_.param("auto_flip", false);
 
-		nh_priv_.param<std::string>("frame_id_prefix", frame_id_prefix_, "aruco_");
+		frame_id_prefix_ = nh_priv_.param<std::string>("frame_id_prefix", "aruco_");
 
 		camera_matrix_ = cv::Mat::zeros(3, 3, CV_64F);
-		dist_coeffs_ = cv::Mat::zeros(8, 1, CV_64F);
 
 		dictionary_ = cv::aruco::getPredefinedDictionary(static_cast<cv::aruco::PREDEFINED_DICTIONARY_NAME>(dictionary));
 		parameters_ = cv::aruco::DetectorParameters::create();
-		parameters_->cornerRefinementMethod = cv::aruco::CORNER_REFINE_SUBPIX;
 
 		image_transport::ImageTransport it(nh_);
 		image_transport::ImageTransport it_priv(nh_priv_);
@@ -170,8 +171,8 @@ private:
 
 				if (!known_tilt_.empty()) {
 					try {
-						snap_to = tf_buffer_.lookupTransform(msg->header.frame_id, known_tilt_,
-						                                     msg->header.stamp, ros::Duration(0.02));
+						snap_to = tf_buffer_->lookupTransform(msg->header.frame_id, known_tilt_,
+						                                      msg->header.stamp, ros::Duration(0.02));
 					} catch (const tf2::TransformException& e) {
 						NODELET_WARN_THROTTLE(5, "can't snap: %s", e.what());
 					}
@@ -205,7 +206,7 @@ private:
 						if (map_markers_ids_.find(ids[i]) == map_markers_ids_.end()) {
 							transform.transform.rotation = marker.pose.orientation;
 							fillTranslation(transform.transform.translation, tvecs[i]);
-							br_.sendTransform(transform);
+							br_->sendTransform(transform);
 						}
 					}
 				}
@@ -326,10 +327,10 @@ private:
 		return frame_id_prefix_ + std::to_string(id);
 	}
 
-	void readLengthOverride()
+	void readLengthOverride(ros::NodeHandle& nh)
 	{
 		std::map<std::string, double> length_override;
-		nh_priv_.getParam("length_override", length_override);
+		nh.getParam("length_override", length_override);
 		for (auto const& item : length_override) {
 			length_override_[std::stoi(item.first)] = item.second;
 		}

aruco_pose/src/aruco_map.cpp

@@ -58,7 +58,6 @@ typedef message_filters::sync_policies::ExactTime<Image, CameraInfo, MarkerArray
 
 class ArucoMap : public nodelet::Nodelet {
 private:
-	ros::NodeHandle nh_, nh_priv_;
 	ros::Publisher img_pub_, pose_pub_, markers_pub_, vis_markers_pub_;
 	image_transport::Publisher debug_pub_;
 	message_filters::Subscriber<Image> image_sub_;
@@ -83,8 +82,8 @@ private:
 public:
 	virtual void onInit()
 	{
-		nh_ = getNodeHandle();
-		nh_priv_ = getPrivateNodeHandle();
+		ros::NodeHandle &nh_ = getNodeHandle();
+		ros::NodeHandle &nh_priv_ = getPrivateNodeHandle();
 
 		image_transport::ImageTransport it_priv(nh_priv_);
 
@@ -96,19 +95,18 @@ public:
 		board_->dictionary = cv::aruco::getPredefinedDictionary(
 			                 static_cast<cv::aruco::PREDEFINED_DICTIONARY_NAME>(nh_priv_.param("dictionary", 2)));
 		camera_matrix_ = cv::Mat::zeros(3, 3, CV_64F);
-		dist_coeffs_ = cv::Mat::zeros(8, 1, CV_64F);
 
 		std::string type, map;
-		nh_priv_.param<std::string>("type", type, "map");
-		nh_priv_.param<std::string>("frame_id", transform_.child_frame_id, "aruco_map");
-		nh_priv_.param<std::string>("known_tilt", known_tilt_, "");
-		nh_priv_.param("auto_flip", auto_flip_, false);
-		nh_priv_.param("image_width", image_width_, 2000);
-		nh_priv_.param("image_height", image_height_, 2000);
-		nh_priv_.param("image_margin", image_margin_, 200);
-		nh_priv_.param("image_axis", image_axis_, true);
-		nh_priv_.param<std::string>("markers/frame_id", markers_parent_frame_, transform_.child_frame_id);
-		nh_priv_.param<std::string>("markers/child_frame_id_prefix", markers_frame_, "");
+		type = nh_priv_.param<std::string>("type", "map");
+		transform_.child_frame_id = nh_priv_.param<std::string>("frame_id", "aruco_map");
+		known_tilt_ = nh_priv_.param<std::string>("known_tilt", "");
+		auto_flip_ = nh_priv_.param("auto_flip", false);
+		image_width_ = nh_priv_.param("image_width" , 2000);
+		image_height_ = nh_priv_.param("image_height", 2000);
+		image_margin_ = nh_priv_.param("image_margin", 200);
+		image_axis_ = nh_priv_.param("image_axis", true);
+		markers_parent_frame_ = nh_priv_.param<std::string>("markers/frame_id", transform_.child_frame_id);
+		markers_frame_ = nh_priv_.param<std::string>("markers/child_frame_id_prefix", "");
 
 		// createStripLine();
 
@@ -116,7 +114,7 @@ public:
 			param(nh_priv_, "map", map);
 			loadMap(map);
 		} else if (type == "gridboard") {
-			createGridBoard();
+			createGridBoard(nh_priv_);
 		} else {
 			NODELET_FATAL("unknown type: %s", type.c_str());
 			ros::shutdown();
@@ -331,7 +329,7 @@ publish_debug:
 		NODELET_INFO("loading %s complete (%d markers)", filename.c_str(), static_cast<int>(board_->ids.size()));
 	}
 
-	void createGridBoard()
+	void createGridBoard(ros::NodeHandle& nh)
 	{
 		NODELET_INFO("generate gridboard");
 		NODELET_WARN("gridboard maps are deprecated");
@@ -339,15 +337,15 @@ publish_debug:
 		int markers_x, markers_y, first_marker;
 		double markers_side, markers_sep_x, markers_sep_y;
 		std::vector<int> marker_ids;
-		nh_priv_.param<int>("markers_x", markers_x, 10);
-		nh_priv_.param<int>("markers_y", markers_y, 10);
-		nh_priv_.param<int>("first_marker", first_marker, 0);
+		markers_x = nh.param("markers_x", 10);
+		markers_y = nh.param("markers_y", 10);
+		first_marker = nh.param("first_marker", 0);
 
-		param(nh_priv_, "markers_side", markers_side);
-		param(nh_priv_, "markers_sep_x", markers_sep_x);
-		param(nh_priv_, "markers_sep_y", markers_sep_y);
+		param(nh, "markers_side", markers_side);
+		param(nh, "markers_sep_x", markers_sep_x);
+		param(nh, "markers_sep_y", markers_sep_y);
 
-		if (nh_priv_.getParam("marker_ids", marker_ids)) {
+		if (nh.getParam("marker_ids", marker_ids)) {
 			if ((unsigned int)(markers_x * markers_y) != marker_ids.size()) {
 				NODELET_FATAL("~marker_ids length should be equal to ~markers_x * ~markers_y");
 				ros::shutdown();

aruco_pose/src/utils.h

@@ -35,9 +35,7 @@ static void parseCameraInfo(const sensor_msgs::CameraInfoConstPtr& cinfo, cv::Ma
 	for (unsigned int i = 0; i < 3; ++i)
 		for (unsigned int j = 0; j < 3; ++j)
 			matrix.at<double>(i, j) = cinfo->K[3 * i + j];
-
-	for (unsigned int k = 0; k < cinfo->D.size(); k++)
-		dist.at<double>(k) = cinfo->D[k];
+	dist = cv::Mat(cinfo->D, true);
 }
 
 inline void rotatePoint(cv::Point3f& p, cv::Point3f origin, float angle)

aruco_pose/test/crash_opencv.py

@@ -0,0 +1,18 @@
+import rospy
+import pytest
+
+from visualization_msgs.msg import MarkerArray as VisMarkerArray
+
+
+@pytest.fixture
+def node():
+    return rospy.init_node('aruco_pose_opencv_crash', anonymous=True)
+
+def test_opencv_crashes_img01(node):
+    rospy.wait_for_message('aruco_detect_01/visualization', VisMarkerArray, timeout=5)
+
+def test_opencv_crashes_img02(node):
+    rospy.wait_for_message('aruco_detect_02/visualization', VisMarkerArray, timeout=5)
+
+def test_opencv_crashes_img03(node):
+    rospy.wait_for_message('aruco_detect_03/visualization', VisMarkerArray, timeout=5)

aruco_pose/test/crash_opencv.test

@@ -0,0 +1,51 @@
+<launch>
+    <arg name="corner_method" default="2"/>
+
+    <node pkg="image_publisher" type="image_publisher" name="imgpub_01" args="$(find aruco_pose)/test/crash_image_01.png">
+        <param name="frame_id" value="main_camera_optical"/>
+        <param name="publish_rate" value="10"/>
+        <param name="camera_info_url" value="file://$(find aruco_pose)/test/camera_info.yaml" />
+    </node>
+
+    <node pkg="image_publisher" type="image_publisher" name="imgpub_02" args="$(find aruco_pose)/test/crash_image_02.png">
+        <param name="frame_id" value="main_camera_optical"/>
+        <param name="publish_rate" value="10"/>
+        <param name="camera_info_url" value="file://$(find aruco_pose)/test/camera_info.yaml" />
+    </node>
+
+    <node pkg="image_publisher" type="image_publisher" name="imgpub_03" args="$(find aruco_pose)/test/crash_image_03.png">
+        <param name="frame_id" value="main_camera_optical"/>
+        <param name="publish_rate" value="10"/>
+        <param name="camera_info_url" value="file://$(find aruco_pose)/test/camera_info.yaml" />
+    </node>
+
+    <node pkg="nodelet" type="nodelet" name="nodelet_manager_01" args="manager"/>
+
+    <node pkg="nodelet" clear_params="true" type="nodelet" name="aruco_detect_01" args="load aruco_pose/aruco_detect nodelet_manager_01">
+        <remap from="image_raw" to="imgpub_01/image_raw"/>
+        <remap from="camera_info" to="imgpub_01/camera_info"/>
+        <param name="length" value="0.33"/>
+        <param name="cornerRefinementMethod" value="$(arg corner_method)"/>
+    </node>
+
+    <node pkg="nodelet" type="nodelet" name="nodelet_manager_02" args="manager"/>
+
+    <node pkg="nodelet" clear_params="true" type="nodelet" name="aruco_detect_02" args="load aruco_pose/aruco_detect nodelet_manager_02">
+        <remap from="image_raw" to="imgpub_02/image_raw"/>
+        <remap from="camera_info" to="imgpub_02/camera_info"/>
+        <param name="length" value="0.33"/>
+        <param name="cornerRefinementMethod" value="$(arg corner_method)"/>
+    </node>
+
+    <node pkg="nodelet" type="nodelet" name="nodelet_manager_03" args="manager"/>
+
+    <node pkg="nodelet" clear_params="true" type="nodelet" name="aruco_detect_03" args="load aruco_pose/aruco_detect nodelet_manager_03">
+        <remap from="image_raw" to="imgpub_03/image_raw"/>
+        <remap from="camera_info" to="imgpub_03/camera_info"/>
+        <param name="length" value="0.33"/>
+        <param name="cornerRefinementMethod" value="$(arg corner_method)"/>
+    </node>
+
+    <param name="test_module" value="$(find aruco_pose)/test/crash_opencv.py"/>
+    <test test-name="crash_opencv" pkg="ros_pytest" type="ros_pytest_runner"/>
+</launch>

aruco_pose/vendor/aruco/src/aruco.cpp

@@ -924,6 +924,8 @@ static void _refineCandidateLines(std::vector<Point>& nContours, std::vector<Poi
 	// calculate the line :: who passes through the grouped points
 	Point3f lines[4];
 	for(int i=0; i<4; i++){
+		// Don't try to "interpolate" single points
+		if (cntPts[i].size() < 2) return;
 		lines[i]=_interpolate2Dline(cntPts[i]);
 	}
 

builder/image-software.sh

@@ -57,6 +57,10 @@ my_travis_retry() {
   return $result
 }
 
+echo_stamp "Increase apt retries"
+
+echo "APT::Acquire::Retries \"3\";" > /etc/apt/apt.conf.d/80-retries
+
 echo_stamp "Install apt keys & repos"
 
 # TODO: This STDOUT consist 'OK'

builder/standalone-install.sh

@@ -1,7 +1,7 @@
 #!/bin/bash
 
 # Perform a "standalone install" in a Docker container
-
+set -e
 # Step 1: Install pip
 apt update
 apt install -y curl

clover/src/optical_flow.cpp

@@ -34,9 +34,7 @@ class OpticalFlow : public nodelet::Nodelet
 {
 public:
 	OpticalFlow():
-		camera_matrix_(3, 3, CV_64F),
-		dist_coeffs_(8, 1, CV_64F),
-		tf_listener_(tf_buffer_)
+		camera_matrix_(3, 3, CV_64F)
 	{}
 
 private:
@@ -52,8 +50,8 @@ private:
 	Mat hann_;
 	Mat prev_, curr_;
 	Mat camera_matrix_, dist_coeffs_;
-	tf2_ros::Buffer tf_buffer_;
-	tf2_ros::TransformListener tf_listener_;
+	std::unique_ptr<tf2_ros::Buffer> tf_buffer_;
+	std::unique_ptr<tf2_ros::TransformListener> tf_listener_;
 	bool calc_flow_gyro_;
 
 	void onInit()
@@ -63,11 +61,14 @@ private:
 		image_transport::ImageTransport it(nh);
 		image_transport::ImageTransport it_priv(nh_priv);
 
-		nh.param<std::string>("mavros/local_position/tf/frame_id", local_frame_id_, "map");
-		nh.param<std::string>("mavros/local_position/tf/child_frame_id", fcu_frame_id_, "base_link");
-		nh_priv.param("roi", roi_px_, 128);
-		nh_priv.param("roi_rad", roi_rad_, 0.0);
-		nh_priv.param("calc_flow_gyro", calc_flow_gyro_, false);
+		tf_buffer_.reset(new tf2_ros::Buffer());
+		tf_listener_.reset(new tf2_ros::TransformListener(*tf_buffer_, nh));
+
+		local_frame_id_ = nh.param<std::string>("mavros/local_position/tf/frame_id", "map");
+		fcu_frame_id_ = nh.param<std::string>("mavros/local_position/tf/child_frame_id", "base_link");
+		roi_px_ = nh_priv.param("roi", 128);
+		roi_rad_ = nh_priv.param("roi_rad", 0.0);
+		calc_flow_gyro_ = nh_priv.param("calc_flow_gyro", false);
 
 		img_sub_ = it.subscribeCamera("image_raw", 1, &OpticalFlow::flow, this);
 		img_pub_ = it_priv.advertise("debug", 1);
@@ -91,9 +92,7 @@ private:
 				camera_matrix_.at<double>(i, j) = cinfo->K[3 * i + j];
 			}
 		}
-		for (int k = 0; k < cinfo->D.size(); k++) {
-			dist_coeffs_.at<double>(k) = cinfo->D[k];
-		}
+		dist_coeffs_ = cv::Mat(cinfo->D, true);
 	}
 
 	void drawFlow(Mat& frame, double x, double y, double quality) const
@@ -186,7 +185,7 @@ private:
 			flow_camera.vector.x = flow_y; // +y means counter-clockwise rotation around Y axis
 			flow_camera.vector.y = -flow_x; // +x means clockwise rotation around X axis
 			try {
-				tf_buffer_.transform(flow_camera, flow_fcu, fcu_frame_id_);
+				tf_buffer_->transform(flow_camera, flow_fcu, fcu_frame_id_);
 			} catch (const tf2::TransformException& e) {
 				// transform is not available yet
 				return;
@@ -200,7 +199,7 @@ private:
 				try {
 					auto flow_gyro_camera = calcFlowGyro(msg->header.frame_id, prev_stamp_, msg->header.stamp);
 					static geometry_msgs::Vector3Stamped flow_gyro_fcu;
-					tf_buffer_.transform(flow_gyro_camera, flow_gyro_fcu, fcu_frame_id_);
+					tf_buffer_->transform(flow_gyro_camera, flow_gyro_fcu, fcu_frame_id_);
 					flow_.integrated_xgyro = flow_gyro_fcu.vector.x;
 					flow_.integrated_ygyro = flow_gyro_fcu.vector.y;
 					flow_.integrated_zgyro = flow_gyro_fcu.vector.z;
@@ -247,8 +246,8 @@ private:
 	geometry_msgs::Vector3Stamped calcFlowGyro(const std::string& frame_id, const ros::Time& prev, const ros::Time& curr)
 	{
 		tf2::Quaternion prev_rot, curr_rot;
-		tf2::fromMsg(tf_buffer_.lookupTransform(frame_id, local_frame_id_, prev).transform.rotation, prev_rot);
-		tf2::fromMsg(tf_buffer_.lookupTransform(frame_id, local_frame_id_, curr, ros::Duration(0.1)).transform.rotation, curr_rot);
+		tf2::fromMsg(tf_buffer_->lookupTransform(frame_id, local_frame_id_, prev).transform.rotation, prev_rot);
+		tf2::fromMsg(tf_buffer_->lookupTransform(frame_id, local_frame_id_, curr, ros::Duration(0.1)).transform.rotation, curr_rot);
 
 		geometry_msgs::Vector3Stamped flow;
 		flow.header.frame_id = frame_id;

docs/en/SUMMARY.md

@@ -11,6 +11,7 @@
   * [Initial setup](setup.md)
   * [Sensor calibration](calibration.md)
   * [RC setup](radio.md)
+    * [Using FS-A8S](rc_flysky_a8s.md)
   * [Flight modes](modes.md)
   * [Power setup](power.md)
   * [Failsafe configuration](failsafe.md)
@@ -41,7 +42,8 @@
   * [Interfacing with a sonar](sonar.md)
   * [Computer vision basics](camera.md)
   * [Using rviz and rqt](rviz.md)
-  * [Software autorun](autolaunch.md)  
+  * [Software autorun](autolaunch.md)
+  * [Using JavaScript](javascript.md)
   * [ROS](ros.md)
   * [MAVROS](mavros.md)
 * Supplementary materials
@@ -87,4 +89,5 @@
   * [Copter Hack 2019](copterhack2019.md)
   * [Copter Hack 2018](copterhack2018.md)
   * [Copter Hack 2017](copterhack2017.md)
+  * [Robocross-2019](robocross2019.md)
   * [Camera calibration (legacy)](camera_calib.md)

docs/en/aruco_map.md

@@ -91,13 +91,6 @@ The marker map adheres to the [ROS coordinate system convention](http://www.ros.
 
 In order to enable vision position estimation you should use the following [PX4 parameters](px4_parameters.md).
 
-If you're using **EKF2** estimator (`SYS_MC_EST_GROUP` parameter is set to `ekf2`), make sure the following is set:
-
-* `EKF2_AID_MASK` should have `vision position fusion` and `vision yaw fusion` flags set.
-* Vision angle observations noise: `EKF2_EVA_NOISE` = 0.1 rad.
-* Vision position observations noise: `EKF2_EVP_NOISE` = 0.1 m.
-* `EKF2_EV_DELAY` = 0.
-
 If you're using **LPE** (`SYS_MC_EST_GROUP` parameter is set to `local_position_estimator,attitude_estimator_q`):
 
 * `LPE_FUSION` should have `vision position` and `land detector` flags set. We suggest unsetting the `baro` flag for indoor flights.
@@ -108,6 +101,13 @@ If you're using **LPE** (`SYS_MC_EST_GROUP` parameter is set to `local_position_
 
 <!-- * Compass should not be fused: `ATT_W_MAG` = 0 -->
 
+If you're using **EKF2** estimator (`SYS_MC_EST_GROUP` parameter is set to `ekf2`), make sure the following is set:
+
+* `EKF2_AID_MASK` should have `vision position fusion` and `vision yaw fusion` flags set.
+* Vision angle observations noise: `EKF2_EVA_NOISE` = 0.1 rad.
+* Vision position observations noise: `EKF2_EVP_NOISE` = 0.1 m.
+* `EKF2_EV_DELAY` = 0.
+
 > **Hint** We recommend using **LPE** for marker-based navigation.
 
 You may use [the `selfcheck.py` utility](selfcheck.md) to check your settings.

docs/en/javascript.md

@@ -0,0 +1,54 @@
+# Work with ROS from browser
+
+Using the [`roslibjs`](http://wiki.ros.org/roslibjs) library it's possible to work with all the ROS resources (topics, services, parameters) from JavaScript code within the browser, which allows creating various interactive web applications for drone.
+
+All the required software is preinstalled in [RPi image](image.md) for Clover.
+
+## Example
+
+An example of a web page, working with `roslib.js`:
+
+```html
+<html>
+	<script src="js/roslib.js"></script>
+	<script type="text/javascript">
+		// Establish roslibjs connection
+		var ros = new ROSLIB.Ros({ url: 'ws://' + location.hostname + ':9090' });
+
+		ros.on('connection', function () {
+			// Connection callback
+			alert('Connected');
+		});
+		
+		// Declare get_telemetry service client
+		var getTelemetry = new ROSLIB.Service({ ros: ros, name : '/get_telemetry', serviceType : 'clover/GetTelemetry' });
+
+		// Call get_telemetry
+		getTelemetry.callService(new ROSLIB.ServiceRequest({ frame_id: 'map' }), function(result) {
+			// Service respond callback
+			alert('Telemetry: ' + JSON.stringify(result));
+		});
+
+		// Subscribe to `/mavros/state` topic
+		var stateSub = new ROSLIB.Topic({ ros : ros, name : '/mavros/state', messageType : 'mavros_msgs/State' });
+		stateSub.subscribe(function(msg) {
+			// Topic message callback
+			console.log('State: ', msg);
+		});
+	</script>
+</html>
+```
+
+[Taking off, landing and all the rest operations](programming.md) can be implemented in a similar way.
+
+The page should be placed in the `/home/pi/catkin_ws/src/clover/clover/www/` directory. After that, it will be available at `http://192.168.11.1/<page_name>.html`. When the page is opened, browser should show an alert with the drone telemetry and constantly print the state of the flight controller to the console.
+
+<img src="../assets/js-ros.png" class="center zoom"/>
+
+See additional information in [`roslibjs` tutorial](http://wiki.ros.org/roslibjs/Tutorials/BasicRosFunctionality).
+
+## Web GCS
+
+See an example of simplified web ground control station on Clover at http://192.168.11.1/gcs.html.
+
+<img src="../assets/web-gcs.png" class="center zoom"/>

docs/en/rc_flysky_a8s.md

@@ -0,0 +1,107 @@
+# Using Flysky FS-A8S
+
+The Flysky FS-A8S receiver is compatible with the Flysky FS-i6 and FS-i6x transmitters. The receiver can output both analog PPM and digital S.Bus/i-Bus signals to the flight controller.
+
+S.Bus is the preferred protocol for the receiver.
+
+## Making a cable
+
+> **Note** You don't need to follow these steps if you already have the right cable; read on to learn [how to bind your transmitter](#rc_bind).
+
+1. Gently remove the yellow wire from the receiver connector. Use sharp tweezers to lift up the plastic wire lock:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/01_remove_cable_fs.png" width=300 class="zoom border" alt="a8s wire removal 1">
+        <img src="../assets/flysky_a8s/02_remove_cable_fs.png" width=300 class="zoom border" alt="a8s wire removal 2">
+    </div>
+
+2. [Pixracer only] Remove the green and brown wires from the 5-pin connector:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/03_remove_cable_pixracer.png" width=300 class="zoom border" alt="pixracer wire removal 1">
+        <img src="../assets/flysky_a8s/04_remove_cable_pixracer.png" width=300 class="zoom border" alt="pixracer wire removal 2">
+    </div>
+
+3. [COEX Pix only] Remove the green wire (or blue if the green one is not present) from the 4-pin connector:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/05_remove_cable_coexpix.png" width=300 class="zoom border" alt="coexpix wire removal 1">
+        <img src="../assets/flysky_a8s/06_remove_cable_coexpix.png" width=300 class="zoom border" alt="coexpix wire removal 2">
+    </div>
+
+4. Use side cutters to cut the Dupont connectors:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/07_wirecuts_1.png" width=300 class="zoom border" alt="wire cutting">
+        <img src="../assets/flysky_a8s/08_wirecuts_2.png" width=300 class="zoom border" alt="wire cuts">
+    </div>
+
+5. Strip and tin 5-7 mm of wire from each side:
+
+    <img src="../assets/flysky_a8s/09_wirecuts_stripped.png" width=300 class="zoom border center" alt="tinning prep">
+
+6. Put heat shrinking tubes on the wires:
+
+    <img src="../assets/flysky_a8s/10_heatshrink.png" width=300 class="zoom border center" alt="shrink tubing">
+
+7. Solder the following wires:
+    * black wire from the receiver connector to the black wire from the flight controller connector;
+    * red wire from the receiver connector to the red wire from the flight controller connector;
+    * white wire from the receiver connector to the white (if you're using Pixracer) or yellow (if you're using COEX Pix) wire from the flight controller connector:
+
+    <img src="../assets/flysky_a8s/11_solder_scheme.png" width=300 class="zoom border center" alt="wire soldering">
+
+8. Put the heat shrinking tubes on the solder joints and heat them:
+
+    <img src="../assets/flysky_a8s/12_heatshrink_heat.png" width=300 class="zoom border center" alt="shrink tube heating">
+
+9. Twist your new cable:
+
+    <img src="../assets/flysky_a8s/13_cable_twist.png" width=300 class="zoom border center" alt="twisted cables">
+
+Connect your receiver to the RC IN port on your flight controller:
+
+<div class="image-group">
+    <img src="../assets/flysky_a8s/14_pixracer_rcin.png" width=300 class="zoom border center" alt="pixracer connection">
+    <img src="../assets/flysky_a8s/14_coexpix_rcin.png" width=300 class="zoom border center" alt="coex pix connection">
+</div>
+
+> **Hint** Double check that you're using the RC IN port on the COEX Pix:
+<img src="../assets/coexpix-bottom.jpg" width=300 class="zoom border center" alt="coex pix pinout">
+
+## Binding your transmitter {#rc_bind}
+
+Do the following to bind your transmitter to the FS-A8S receiver:
+
+1. Make sure your flight controller is powered off.
+2. Hold the **BIND** button on the receiver:
+
+    <img src="../assets/flysky_a8s/15_bind_key.png" width=300 class="zoom border center" alt="bind key">
+
+3. Turn on the flight controller. The LED light on the receiver should blink fast, about 3 times per second.
+
+    <img src="../assets/flysky_a8s/16_blink_fast.gif" width=300 class="zoom border center" alt="fast blink">
+
+4. Hold the **BIND KEY** on your transmitter and power it on. You should see a message saying **RX Binding...**
+
+    <img src="../assets/flysky_a8s/17_controller_rxbind.png" width=300 class="zoom border center" alt="rx binding">
+
+5. The LED light on the receiver should start blinking slowly, about once per second.
+
+    <img src="../assets/flysky_a8s/16_blink_slow.gif" width=300 class="zoom border center" alt="slow blink">
+
+6. Turn your transmitter off and on again. The LED light on the receiver should glow steadily.
+
+    <img src="../assets/flysky_a8s/16_bind_indicator.png" width=300 class="zoom border center" alt="steady glow">
+
+> **Note** This receiver does not send any telemetry data back to the transmitter. Your transmitter will not display any data like RSSI and drone battery level on its screen. In fact, there will be no indication that the transmitter is connected to the receiver. This is not a malfunction, the controls will still work.
+
+## Changing the receiver mode (S.Bus/i-Bus)
+
+Connect your flight controller to your computer and open QGroundControl. In it, open the Radio tab:
+
+![qgc radio pane](../assets/flysky_a8s/18_qgc_radio.png)
+
+If it shows zero channels under the transmitter image, hold the **BIND** key on the receiver for 2 seconds. You should then see 18 channels appear under the image:
+
+![qgc radio with channels](../assets/flysky_a8s/19_qgc_channels.png)

docs/en/robocross2019.md

@@ -0,0 +1,25 @@
+# Robocross-2019
+
+On July, 2019, for the fourth time in a row, the team Copter Express won the annual tests of unmanned vehicles "[Robocross](http://russianrobotics.ru/activities/robokross-2019/)". Tests are held at the GAZ test site near Nizhny Novgorod.
+
+The main objective of the tests in the UAV category was to localize and destroy the target - the red balloon - autonomously.
+
+## Video
+
+<iframe width="560" height="315" src="https://www.youtube.com/embed/zMh5THdHuX8" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
+
+## Implementation
+
+The team used an F450 frame based quadcopter and [Clover software platform](https://github.com/CopterExpress/clover). The final source code is available [on GitHub](https://github.com/CopterExpress/robocross2019/).
+
+`robocross2019` ROS package is divided into two parts: `red_dead_detection` ROS nodelet recognizes the red ball, `ball.py` implements high-level flight logic.
+
+## red_dead_detection
+
+The `red_dead_detection` nodelet recognizes the red ball on the image from the forward looking quadcopter camera (`/front_camera/image_raw` and `/front_camera/camera_info` topics). The simplest method of filtering the image by color is applied. Then the nodelet calculates the geometric center of the detected segments, and performs camera distortion compensation (`cv::undistortPoints`).
+
+Using the known focal lengths of the camera (from `camera_info`), the nodelet calculates the vector directed towards the target. The resulting vector is published to the topic `/red_dead_detection/direction`; its coordinate system (`frame_id` is associated with the front camera `front_camera_optical`).
+
+## balloon.py
+
+To fly towards the ball, the direction vector `red_dead_detection/direction` is used, which is set as a setpoint for the velocity of the drone. The yaw angle is also set towards the ball. The target is considered destroyed when the total area of red pixels is less than the threshold for a certain amount of camera frames.

docs/ru/SUMMARY.md

@@ -11,6 +11,7 @@
   * [Первоначальная настройка](setup.md)
   * [Калибровка датчиков](calibration.md)
   * [Настройка пульта](radio.md)
+    * [Работа с FS-A8S](rc_flysky_a8s.md)
   * [Полетные режимы](modes.md)
   * [Настройка питания](power.md)
   * [Настройка failsafe](failsafe.md)
@@ -42,6 +43,7 @@
   * [Компьютерное зрение](camera.md)
   * [Визуализация с помощью rviz](rviz.md)
   * [Автозапуск ПО](autolaunch.md)
+  * [Использование JavaScript](javascript.md)
   * [ROS](ros.md)
   * [MAVROS](mavros.md)
 * Дополнительные материалы

docs/ru/aruco_map.md

@@ -91,23 +91,23 @@ rosrun aruco_pose genmap.py 0.33 2 4 1 1 0 > ~/catkin_ws/src/clover/aruco_pose/m
 
 Для работы механизма Vision Position Estimation необходимы следующие [настройки PX4](px4_parameters.md).
 
-При использовании **EKF2** (параметр `SYS_MC_EST_GROUP` = `ekf2`):
-
-* В параметре `EKF2_AID_MASK` включены флажки `vision position fusion`, `vision yaw fusion`.
-* Шум угла по зрению: `EKF2_EVA_NOISE` = 0.1 rad
-* Шум позиции по зрению: `EKF2_EVP_NOISE` = 0.1 m
-* `EKF2_EV_DELAY` = 0
-
 При использовании **LPE** (параметр `SYS_MC_EST_GROUP` = `local_position_estimator, attitude_estimator_q`):
 
 * В параметре `LPE_FUSION` включены флажки `vision position`, `land detector`. Флажок `baro` рекомендуется отключить.
 * Вес угла по рысканью по зрению: `ATT_W_EXT_HDG` = 0.5
 * Включена ориентация по Yaw по зрению: `ATT_EXT_HDG_M` = 1 `Vision`.
 * Шумы позиции по зрению: `LPE_VIS_XY` = 0.1 m, `LPE_VIS_Z` = 0.1 m.
-* `LPE_VIS_DELAY` = 0 sec
+* `LPE_VIS_DELAY` = 0 sec.
 
 <!-- * Выключен компас: `ATT_W_MAG` = 0 -->
 
+При использовании **EKF2** (параметр `SYS_MC_EST_GROUP` = `ekf2`):
+
+* В параметре `EKF2_AID_MASK` включены флажки `vision position fusion`, `vision yaw fusion`.
+* Шум угла по зрению: `EKF2_EVA_NOISE` = 0.1 rad.
+* Шум позиции по зрению: `EKF2_EVP_NOISE` = 0.1 m.
+* `EKF2_EV_DELAY` = 0.
+
 > **Hint** На данный момент для полета по маркерам рекомендуется использование **LPE**.
 
 Для проверки правильности всех настроек можно [воспользоваться утилитой `selfcheck.py`](selfcheck.md).

docs/ru/javascript.md

@@ -0,0 +1,54 @@
+# Работа с ROS из браузера
+
+С помощью библиотеки [`roslibjs`](http://wiki.ros.org/roslibjs) возможна работа со всеми ресурсами ROS (топики, сервисы, параметры) из JavaScript-кода внутри браузера, что позволяет создавать различные интерактивные браузерные приложения для коптера.
+
+Все необходимое для работы с `roslibjs` предустановлено и настроено на [образе для RPi для Клевера](image.md).
+
+## Пример
+
+Пример HTML-кода страницы, работающей с `roslib.js`:
+
+```html
+<html>
+	<script src="js/roslib.js"></script>
+	<script type="text/javascript">
+		// Establish roslibjs connection
+		var ros = new ROSLIB.Ros({ url: 'ws://' + location.hostname + ':9090' });
+
+		ros.on('connection', function () {
+			// Connection callback
+			alert('Connected');
+		});
+		
+		// Declare get_telemetry service client
+		var getTelemetry = new ROSLIB.Service({ ros: ros, name : '/get_telemetry', serviceType : 'clover/GetTelemetry' });
+
+		// Call get_telemetry
+		getTelemetry.callService(new ROSLIB.ServiceRequest({ frame_id: 'map' }), function(result) {
+			// Service respond callback
+			alert('Telemetry: ' + JSON.stringify(result));
+		});
+
+		// Subscribe to `/mavros/state` topic
+		var stateSub = new ROSLIB.Topic({ ros : ros, name : '/mavros/state', messageType : 'mavros_msgs/State' });
+		stateSub.subscribe(function(msg) {
+			// Topic message callback
+			console.log('State: ', msg);
+		});
+	</script>
+</html>
+```
+
+[Взлет, посадка и все остальные операции](programming.md) могут быть осуществлены аналогичным образом.
+
+Страница должна быть помещена в каталог `/home/pi/catkin_ws/src/clover/clover/www/`. После этого она станет доступна по адресу `http://192.168.11.1/<имя_страницы>.html`. При открытии страницы браузер должен показать окно с телеметрией дрона, а также постоянно выводить состояние полетного контроллера в консоль.
+
+<img src="../assets/js-ros.png" class="center zoom"/>
+
+Более подробную информацию смотрите в [туториале по `roslibjs`](http://wiki.ros.org/roslibjs/Tutorials/BasicRosFunctionality).
+
+## Браузерная GCS
+
+Смотрите также пример реализации упрощенной браузерной наземной станции (GCS) на Клевере по адресу http://192.168.11.1/gcs.html.
+
+<img src="../assets/web-gcs.png" class="center zoom"/>

docs/ru/rc_flysky_a8s.md

@@ -0,0 +1,107 @@
+# Работа с приёмником Flysky FS-A8S
+
+Приёмник Flysky FS-A8S совместим с пультами Flysky FS-i6 и FS-i6x. Связь с полётным контроллером может происходить как с использованием аналогового протокола PPM, так и при помощи цифровых протоколов S.Bus/i-Bus.
+
+Для подключения к полётному контроллеру рекомендуется использовать протокол S.Bus.
+
+## Изготовление кабеля для подключения к полётному контроллеру
+
+> **Note** Если в вашем наборе уже есть кабель для подключения приёмника к полётному контроллеру, совместимому с вашим, переходите к [следующему этапу](#rc_bind).
+
+1. Аккуратно извлеките жёлтый провод из коннектора, идущего к приёмнику. Для того, чтобы вытащить провод, приподнимите острым пинцетом замок коннектора:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/01_remove_cable_fs.png" width=300 class="zoom border" alt="a8s wire removal 1">
+        <img src="../assets/flysky_a8s/02_remove_cable_fs.png" width=300 class="zoom border" alt="a8s wire removal 2">
+    </div>
+
+2. [При использовании полётного контроллера Pixracer] Извлеките 2 провода - зелёный и коричневый - из 5-пинового коннектора для полётного контроллера:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/03_remove_cable_pixracer.png" width=300 class="zoom border" alt="pixracer wire removal 1">
+        <img src="../assets/flysky_a8s/04_remove_cable_pixracer.png" width=300 class="zoom border" alt="pixracer wire removal 2">
+    </div>
+
+3. [При использовании полётного контроллера COEX Pix] Извлеките зелёный (или синий, в зависимости от комплектации) провод из 4-пинового коннектора для полётного контроллера:
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/05_remove_cable_coexpix.png" width=300 class="zoom border" alt="coexpix wire removal 1">
+        <img src="../assets/flysky_a8s/06_remove_cable_coexpix.png" width=300 class="zoom border" alt="coexpix wire removal 2">
+    </div>
+
+4. С помощью бокорезов обрежьте концы кабелей с разъёмами Dupont (чёрными):
+
+    <div class="image-group">
+        <img src="../assets/flysky_a8s/07_wirecuts_1.png" width=300 class="zoom border" alt="wire cutting">
+        <img src="../assets/flysky_a8s/08_wirecuts_2.png" width=300 class="zoom border" alt="wire cuts">
+    </div>
+
+5. Зачистите и залудите 5-7 мм провода с каждой стороны:
+
+    <img src="../assets/flysky_a8s/09_wirecuts_stripped.png" width=300 class="zoom border center" alt="tinning prep">
+
+6. Наденьте термоусадочные трубки на провода:
+
+    <img src="../assets/flysky_a8s/10_heatshrink.png" width=300 class="zoom border center" alt="shrink tubing">
+
+7. Спаяйте провода по следующей схеме:
+    * чёрный провод приёмника с чёрным проводом кабеля полётного контроллера;
+    * красный провод приёмника с красным проводом кабеля полётного контроллера;
+    * белый провод приёмника с белым (при использовании Pixracer) или жёлтым (при использовании COEX Pix) проводом кабеля полётного контроллера:
+
+    <img src="../assets/flysky_a8s/11_solder_scheme.png" width=300 class="zoom border center" alt="wire soldering">
+
+8. Переместите термоусадочные трубки на места пайки и прогрейте их феном:
+
+    <img src="../assets/flysky_a8s/12_heatshrink_heat.png" width=300 class="zoom border center" alt="shrink tube heating">
+
+9. Скрутите готовые кабели:
+
+    <img src="../assets/flysky_a8s/13_cable_twist.png" width=300 class="zoom border center" alt="twisted cables">
+
+С помощью изготовленного кабеля подключите приёмник к полётному контроллеру к порту RC IN:
+
+<div class="image-group">
+    <img src="../assets/flysky_a8s/14_pixracer_rcin.png" width=300 class="zoom border center" alt="pixracer connection">
+    <img src="../assets/flysky_a8s/14_coexpix_rcin.png" width=300 class="zoom border center" alt="coex pix connection">
+</div>
+
+> **Hint** Убедитесь, что провод, идущий в COEX Pix, подключен к порту RC IN:
+<img src="../assets/coexpix-bottom.jpg" width=300 class="zoom border center" alt="coex pix pinout">
+
+## Сопряжение приёмника с пультом {#rc_bind}
+
+Для сопряжения приёмника с пультом:
+
+1. Убедитесь, что полётный контроллер выключен.
+2. Зажмите кнопку **BIND** на приёмнике:
+
+    <img src="../assets/flysky_a8s/15_bind_key.png" width=300 class="zoom border center" alt="bind key">
+
+3. Включите полётный контроллер. Светодиод на приёмнике должен замигать с высокой частотой.
+
+    <img src="../assets/flysky_a8s/16_blink_fast.gif" width=300 class="zoom border center" alt="fast blink">
+
+4. Зажмите клавишу **BIND KEY** на пульте и включите его. На пульте должно появиться сообщение **RX Binding...**
+
+    <img src="../assets/flysky_a8s/17_controller_rxbind.png" width=300 class="zoom border center" alt="rx binding">
+
+5. Светодиод на приёмнике должен замигать с низкой частотой.
+
+    <img src="../assets/flysky_a8s/16_blink_slow.gif" width=300 class="zoom border center" alt="slow blink">
+
+6. Выключите и включите пульт. Светодиод на приёмнике должен светиться непрерывно.
+
+    <img src="../assets/flysky_a8s/16_bind_indicator.png" width=300 class="zoom border center" alt="steady glow">
+
+> **Note** Данный приёмник не передаёт телеметрию обратно на пульт. Это значит, что на экране пульта не будет отображаться информация об уровне сигнала и напряжении на приёмнике. Эта особенность приёмника не является неисправностью и не влияет на управление.
+
+## Выбор режима приёмника
+
+Откройте QGroundControl и подключите полётный контроллер к компьютеру. Откройте вкладку Radio:
+
+![qgc radio pane](../assets/flysky_a8s/18_qgc_radio.png)
+
+Если справа (под изображением пульта) не показано ни одного канала, зажмите кнопку **BIND** на приёмнике на 2 секунды. Должно появиться 18 каналов:
+
+![qgc radio with channels](../assets/flysky_a8s/19_qgc_channels.png)