Point Cloud Library (PCL) 1.12.1
conversions.h
1/*
2 * Software License Agreement (BSD License)
3 *
4 * Point Cloud Library (PCL) - www.pointclouds.org
5 * Copyright (c) 2010-2012, Willow Garage, Inc.
6 *
7 * All rights reserved.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 *
13 * * Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * * Redistributions in binary form must reproduce the above
16 * copyright notice, this list of conditions and the following
17 * disclaimer in the documentation and/or other materials provided
18 * with the distribution.
19 * * Neither the name of the copyright holder(s) nor the names of its
20 * contributors may be used to endorse or promote products derived
21 * from this software without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
24 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
25 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
26 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
27 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
28 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
29 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
30 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
31 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
33 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
34 * POSSIBILITY OF SUCH DAMAGE.
35 *
36 * $Id$
37 *
38 */
39
40#pragma once
41
42#ifdef __GNUC__
43#pragma GCC system_header
44#endif
45
46#include <pcl/PCLPointField.h>
47#include <pcl/PCLPointCloud2.h>
48#include <pcl/PCLImage.h>
49#include <pcl/point_cloud.h>
50#include <pcl/type_traits.h>
51#include <pcl/for_each_type.h>
52#include <pcl/console/print.h>
53
54#include <boost/foreach.hpp>
55
56namespace pcl
57{
58 namespace detail
59 {
60 // For converting template point cloud to message.
61 template<typename PointT>
63 {
64 FieldAdder (std::vector<pcl::PCLPointField>& fields) : fields_ (fields) {};
65
66 template<typename U> void operator() ()
67 {
69 f.name = pcl::traits::name<PointT, U>::value;
70 f.offset = pcl::traits::offset<PointT, U>::value;
71 f.datatype = pcl::traits::datatype<PointT, U>::value;
72 f.count = pcl::traits::datatype<PointT, U>::size;
73 fields_.push_back (f);
74 }
75
76 std::vector<pcl::PCLPointField>& fields_;
77 };
78
79 // For converting message to template point cloud.
80 template<typename PointT>
82 {
83 FieldMapper (const std::vector<pcl::PCLPointField>& fields,
84 std::vector<FieldMapping>& map)
85 : fields_ (fields), map_ (map)
86 {
87 }
88
89 template<typename Tag> void
91 {
92 for (const auto& field : fields_)
93 {
94 if (FieldMatches<PointT, Tag>()(field))
95 {
96 FieldMapping mapping;
97 mapping.serialized_offset = field.offset;
98 mapping.struct_offset = pcl::traits::offset<PointT, Tag>::value;
99 mapping.size = sizeof (typename pcl::traits::datatype<PointT, Tag>::type);
100 map_.push_back (mapping);
101 return;
102 }
103 }
104 // Disable thrown exception per #595: http://dev.pointclouds.org/issues/595
105 PCL_WARN ("Failed to find match for field '%s'.\n", pcl::traits::name<PointT, Tag>::value);
106 //throw pcl::InvalidConversionException (ss.str ());
107 }
108
109 const std::vector<pcl::PCLPointField>& fields_;
110 std::vector<FieldMapping>& map_;
111 };
112
113 inline bool
115 {
117 }
118
119 } //namespace detail
120
121 template<typename PointT> void
122 createMapping (const std::vector<pcl::PCLPointField>& msg_fields, MsgFieldMap& field_map)
123 {
124 // Create initial 1-1 mapping between serialized data segments and struct fields
125 detail::FieldMapper<PointT> mapper (msg_fields, field_map);
126 for_each_type< typename traits::fieldList<PointT>::type > (mapper);
127
128 // Coalesce adjacent fields into single memcpy's where possible
129 if (field_map.size() > 1)
130 {
131 std::sort(field_map.begin(), field_map.end(), detail::fieldOrdering);
132 MsgFieldMap::iterator i = field_map.begin(), j = i + 1;
133 while (j != field_map.end())
134 {
135 // This check is designed to permit padding between adjacent fields.
136 /// @todo One could construct a pathological case where the struct has a
137 /// field where the serialized data has padding
138 if (j->serialized_offset - i->serialized_offset == j->struct_offset - i->struct_offset)
139 {
140 i->size += (j->struct_offset + j->size) - (i->struct_offset + i->size);
141 j = field_map.erase(j);
142 }
143 else
144 {
145 ++i;
146 ++j;
147 }
148 }
149 }
150 }
151
152 /** \brief Convert a PCLPointCloud2 binary data blob into a pcl::PointCloud<T> object using a field_map.
153 * \param[in] msg the PCLPointCloud2 binary blob
154 * \param[out] cloud the resultant pcl::PointCloud<T>
155 * \param[in] field_map a MsgFieldMap object
156 *
157 * \note Use fromPCLPointCloud2 (PCLPointCloud2, PointCloud<T>) directly or create you
158 * own MsgFieldMap using:
159 *
160 * \code
161 * MsgFieldMap field_map;
162 * createMapping<PointT> (msg.fields, field_map);
163 * \endcode
164 */
165 template <typename PointT> void
167 const MsgFieldMap& field_map)
168 {
169 // Copy info fields
170 cloud.header = msg.header;
171 cloud.width = msg.width;
172 cloud.height = msg.height;
173 cloud.is_dense = msg.is_dense == 1;
174
175 // Copy point data
176 cloud.resize (msg.width * msg.height);
177 std::uint8_t* cloud_data = reinterpret_cast<std::uint8_t*>(&cloud[0]);
178
179 // Check if we can copy adjacent points in a single memcpy. We can do so if there
180 // is exactly one field to copy and it is the same size as the source and destination
181 // point types.
182 if (field_map.size() == 1 &&
183 field_map[0].serialized_offset == 0 &&
184 field_map[0].struct_offset == 0 &&
185 field_map[0].size == msg.point_step &&
186 field_map[0].size == sizeof(PointT))
187 {
188 const auto cloud_row_step = (sizeof (PointT) * cloud.width);
189 const std::uint8_t* msg_data = &msg.data[0];
190 // Should usually be able to copy all rows at once
191 if (msg.row_step == cloud_row_step)
192 {
193 memcpy (cloud_data, msg_data, msg.data.size ());
194 }
195 else
196 {
197 for (uindex_t i = 0; i < msg.height; ++i, cloud_data += cloud_row_step, msg_data += msg.row_step)
198 memcpy (cloud_data, msg_data, cloud_row_step);
199 }
200
201 }
202 else
203 {
204 // If not, memcpy each group of contiguous fields separately
205 for (uindex_t row = 0; row < msg.height; ++row)
206 {
207 const std::uint8_t* row_data = &msg.data[row * msg.row_step];
208 for (uindex_t col = 0; col < msg.width; ++col)
209 {
210 const std::uint8_t* msg_data = row_data + col * msg.point_step;
211 for (const detail::FieldMapping& mapping : field_map)
212 {
213 memcpy (cloud_data + mapping.struct_offset, msg_data + mapping.serialized_offset, mapping.size);
214 }
215 cloud_data += sizeof (PointT);
216 }
217 }
218 }
219 }
220
221 /** \brief Convert a PCLPointCloud2 binary data blob into a pcl::PointCloud<T> object.
222 * \param[in] msg the PCLPointCloud2 binary blob
223 * \param[out] cloud the resultant pcl::PointCloud<T>
224 */
225 template<typename PointT> void
227 {
228 MsgFieldMap field_map;
229 createMapping<PointT> (msg.fields, field_map);
230 fromPCLPointCloud2 (msg, cloud, field_map);
231 }
232
233 /** \brief Convert a pcl::PointCloud<T> object to a PCLPointCloud2 binary data blob.
234 * \param[in] cloud the input pcl::PointCloud<T>
235 * \param[out] msg the resultant PCLPointCloud2 binary blob
236 */
237 template<typename PointT> void
239 {
240 // Ease the user's burden on specifying width/height for unorganized datasets
241 if (cloud.width == 0 && cloud.height == 0)
242 {
243 msg.width = cloud.size ();
244 msg.height = 1;
245 }
246 else
247 {
248 assert (cloud.size () == cloud.width * cloud.height);
249 msg.height = cloud.height;
250 msg.width = cloud.width;
251 }
252
253 // Fill point cloud binary data (padding and all)
254 std::size_t data_size = sizeof (PointT) * cloud.size ();
255 msg.data.resize (data_size);
256 if (data_size)
257 {
258 memcpy(&msg.data[0], &cloud[0], data_size);
259 }
260
261 // Fill fields metadata
262 msg.fields.clear ();
263 for_each_type<typename traits::fieldList<PointT>::type> (detail::FieldAdder<PointT>(msg.fields));
264
265 msg.header = cloud.header;
266 msg.point_step = sizeof (PointT);
267 msg.row_step = (sizeof (PointT) * msg.width);
268 msg.is_dense = cloud.is_dense;
269 /// @todo msg.is_bigendian = ?;
270 }
271
272 /** \brief Copy the RGB fields of a PointCloud into pcl::PCLImage format
273 * \param[in] cloud the point cloud message
274 * \param[out] msg the resultant pcl::PCLImage
275 * CloudT cloud type, CloudT should be akin to pcl::PointCloud<pcl::PointXYZRGBA>
276 * \note will throw std::runtime_error if there is a problem
277 */
278 template<typename CloudT> void
279 toPCLPointCloud2 (const CloudT& cloud, pcl::PCLImage& msg)
280 {
281 // Ease the user's burden on specifying width/height for unorganized datasets
282 if (cloud.width == 0 && cloud.height == 0)
283 throw std::runtime_error("Needs to be a dense like cloud!!");
284 else
285 {
286 if (cloud.size () != cloud.width * cloud.height)
287 throw std::runtime_error("The width and height do not match the cloud size!");
288 msg.height = cloud.height;
289 msg.width = cloud.width;
290 }
291
292 // ensor_msgs::image_encodings::BGR8;
293 msg.header = cloud.header;
294 msg.encoding = "bgr8";
295 msg.step = msg.width * sizeof (std::uint8_t) * 3;
296 msg.data.resize (msg.step * msg.height);
297 for (std::size_t y = 0; y < cloud.height; y++)
298 {
299 for (std::size_t x = 0; x < cloud.width; x++)
300 {
301 std::uint8_t * pixel = &(msg.data[y * msg.step + x * 3]);
302 memcpy (pixel, &cloud (x, y).rgb, 3 * sizeof(std::uint8_t));
303 }
304 }
305 }
306
307 /** \brief Copy the RGB fields of a PCLPointCloud2 msg into pcl::PCLImage format
308 * \param cloud the point cloud message
309 * \param msg the resultant pcl::PCLImage
310 * will throw std::runtime_error if there is a problem
311 */
312 inline void
314 {
315 const auto predicate = [](const auto& field) { return field.name == "rgb"; };
316 const auto result = std::find_if(cloud.fields.cbegin (), cloud.fields.cend (), predicate);
317 if (result == cloud.fields.end ())
318 throw std::runtime_error ("No rgb field!!");
319
320 const auto rgb_index = std::distance(cloud.fields.begin (), result);
321 if (cloud.width == 0 && cloud.height == 0)
322 throw std::runtime_error ("Needs to be a dense like cloud!!");
323 else
324 {
325 msg.height = cloud.height;
326 msg.width = cloud.width;
327 }
328 auto rgb_offset = cloud.fields[rgb_index].offset;
329 const auto point_step = cloud.point_step;
330
331 // pcl::image_encodings::BGR8;
332 msg.header = cloud.header;
333 msg.encoding = "bgr8";
334 msg.step = (msg.width * sizeof (std::uint8_t) * 3);
335 msg.data.resize (msg.step * msg.height);
336
337 for (std::size_t y = 0; y < cloud.height; y++)
338 {
339 for (std::size_t x = 0; x < cloud.width; x++, rgb_offset += point_step)
340 {
341 std::uint8_t * pixel = &(msg.data[y * msg.step + x * 3]);
342 memcpy (pixel, &(cloud.data[rgb_offset]), 3 * sizeof (std::uint8_t));
343 }
344 }
345 }
346}
PointCloud represents the base class in PCL for storing collections of 3D points.
Definition: point_cloud.h:173
bool is_dense
True if no points are invalid (e.g., have NaN or Inf values in any of their floating point fields).
Definition: point_cloud.h:403
void resize(std::size_t count)
Resizes the container to contain count elements.
Definition: point_cloud.h:462
std::uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:398
pcl::PCLHeader header
The point cloud header.
Definition: point_cloud.h:392
std::uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:400
std::size_t size() const
Definition: point_cloud.h:443
bool fieldOrdering(const FieldMapping &a, const FieldMapping &b)
Definition: conversions.h:114
float distance(const PointT &p1, const PointT &p2)
Definition: geometry.h:60
detail::int_type_t< detail::index_type_size, false > uindex_t
Type used for an unsigned index in PCL.
Definition: types.h:120
void toPCLPointCloud2(const pcl::PointCloud< PointT > &cloud, pcl::PCLPointCloud2 &msg)
Convert a pcl::PointCloud<T> object to a PCLPointCloud2 binary data blob.
Definition: conversions.h:238
void createMapping(const std::vector< pcl::PCLPointField > &msg_fields, MsgFieldMap &field_map)
Definition: conversions.h:122
void fromPCLPointCloud2(const pcl::PCLPointCloud2 &msg, pcl::PointCloud< PointT > &cloud, const MsgFieldMap &field_map)
Convert a PCLPointCloud2 binary data blob into a pcl::PointCloud<T> object using a field_map.
Definition: conversions.h:166
std::vector< detail::FieldMapping > MsgFieldMap
Definition: point_cloud.h:72
uindex_t step
Definition: PCLImage.h:21
uindex_t height
Definition: PCLImage.h:16
std::string encoding
Definition: PCLImage.h:18
std::vector< std::uint8_t > data
Definition: PCLImage.h:23
uindex_t width
Definition: PCLImage.h:17
::pcl::PCLHeader header
Definition: PCLImage.h:14
std::uint8_t is_dense
std::vector<::pcl::PCLPointField > fields
::pcl::PCLHeader header
std::vector< std::uint8_t > data
std::string name
Definition: PCLPointField.h:14
std::uint8_t datatype
Definition: PCLPointField.h:17
A point structure representing Euclidean xyz coordinates, and the RGB color.
FieldAdder(std::vector< pcl::PCLPointField > &fields)
Definition: conversions.h:64
std::vector< pcl::PCLPointField > & fields_
Definition: conversions.h:76
FieldMapper(const std::vector< pcl::PCLPointField > &fields, std::vector< FieldMapping > &map)
Definition: conversions.h:83
const std::vector< pcl::PCLPointField > & fields_
Definition: conversions.h:109
std::vector< FieldMapping > & map_
Definition: conversions.h:110
std::size_t serialized_offset
Definition: point_cloud.h:64