memo

+      cv::fillConvexPoly(points_img, approx, approx.size(), 255);
+      cv::dilate(points_img, points_img, cv::Mat());
+      point_list.clear();
+      for (int i=0;i<points_img.cols;i++){
+        for (int j=0;j<points_img.rows;j++){
+          if (points_img.at<uchar>(j,i) == 255)
+          {
+            cv::Point p;
+            p.x = i;
+            p.y = j;
+            point_list.push_back(p);
+          }
+        }
+      }
+      approx.clear();
+      cv::convexHull(point_list, approx);

docker push やり方

https://zenn.dev/katan/articles/1d5ff92fd809e7

shape_based_matching : docker repo https://hub.docker.com/repository/docker/hoshianaaa222/shape_based_mathcing

docker ps
docker commit 090bdc6ffd84 hoshianaaa222/shape_based_matching:latest
docker push hoshianaaa222/shape_based_matching:latest

エラー

info.angle: 0
too few features, abort
OpenCV Error: Assertion failed (i->second.size() > size_t(template_id)) in getTemplates, file /shape_based_matching/line2Dup.cpp, line 1367
terminate called after throwing an instance of 'cv::Exception'
  what():  /shape_based_matching/line2Dup.cpp:1367: error: (-215) i->second.size() > size_t(template_id) in function getTemplates

angle_train関数内部に以下を追加

try
{

}
catch(...)
{

}

angle_test error

• solution: fix file path

before

after

shape_based_matching

update:
fusion implementation to run faster!
icp is also refined to be faster and easier to use

Transforms in shape-based matching
pose refine with icp branch, 0.1-0.5 degree accuracy
icp + subpixel branch, < 0.1 degree accuracy
icp + subpixel + sim3(previous is so3) branch, deal with scale error

try to implement halcon shape based matching, refer to machine vision algorithms and applications, page 317 3.11.5, written by halcon engineers
We find that shape based matching is the same as linemod. linemod pdf

halcon match solution guide for how to select matching methods(halcon documentation):

steps

1. change test.cpp line 9 prefix to top level folder

2. in cmakeList line 23, change /opt/ros/kinetic to somewhere opencv3 can be found(if opencv3 is installed in default env then don't need to)

3. cmake make & run. To learn usage, see different tests in test.cpp. Particularly, scale_test are fully commented.

NOTE: On windows, it's confirmed that visual studio 17 works fine, but there are some problems with MIPP in vs13. You may want old codes without MIPP: old commit

thoughts about the method

The key of shape based matching, or linemod, is using gradient orientation only. Though both edge and orientation are resistant to disturbance, edge have only 1bit info(there is an edge or not), so it's hard to dig wanted shapes out if there are too many edges, but we have to have as many edges as possible if we want to find all the target shapes. It's quite a dilemma.

However, gradient orientation has much more info than edge, so we can easily match shape orientation in the overwhelming img orientation by template matching across the img.

Speed is also important. Thanks to the speeding up magic in linemod, we can handle 1000 templates in 20ms or so.

Chinese blog about the thoughts

improvment

Comparing to opencv linemod src, we improve from 6 aspects:

1. delete depth modality so we don't need virtual func, this may speed up

2. opencv linemod can't use more than 63 features. Now wo can have up to 8191

3. simple codes for rotating and scaling img for training. see test.cpp for examples

4. nms for accurate edge selection

5. one channel orientation extraction to save time, slightly faster for gray img

6. use MIPP for multiple platforms SIMD, for example, x86 SSE AVX, arm neon. To have better performance, we have extended MIPP to uint8_t for some instructions.(Otherwise we can only use half feature points to avoid int8_t overflow)

7. rotate features directly to speed up template extractions; selectScatteredFeatures more evenly; exautive select all features if not enough rather than abort templates(but features <= 4 will abort)

some test

Example for circle shape

You can imagine how many circles we will find if use edges

Not that circular

Blur

circle template before and after nms

before nms

after nms

Simple example for arbitary shape

Well, the example is too simple to show the robustness
running time: 1024x1024, 60ms to construct response map, 7ms for 360 templates

test img & templ features

noise test

some issues you may want to know

Well, issues are not clearly classified and many questions are discussed in one issue sometimes. For better reference, some typical discussions are pasted here.

object too small?
failure case?
how to run even faster?

このrepoを動かしてみる

https://github.com/meiqua/shape_based_matching

• 1. docker kinetic image 実行

docker ros kinetic https://qiita.com/seigot/items/9a3a101dbbd7828ee00b

docker run -it --rm -p 6080:80 ct2034/vnc-ros-kinetic-full
docker run -it --rm -p 6080:80 ct2034/vnc-ros-kinetic-full/shape-based-matching:0.0.3

• commit

docker commit {ps id} ct2034/vnc-ros-kinetic-full/shape-based-matching:0.0.3

ryodoさんの https://github.com/RyodoTanaka/ubuntu

http://127.0.0.1:6080/ にアクセス

• 2. その中にcloneしてtest.cppのテストフォルダへのパスを修正　ビルド&実行
• 3. 結果

test.cppに3種類のサンプルがあるみたい

上の写真は,angle_test()関数

• scale_test()実行結果

• noise_test()実行結果

• angle_test() : モードを"test"じゃなくしてみる ※ 引数を ("test", True)　から ("train", True)に変更

結果

• scale_test()でモードをtestからtrainに変更

以下のような出力がされた

empl_id: 0
templ_id: 1
have no enough features, exaustive mode
templ_id: 2
have no enough features, exaustive mode
templ_id: 3
templ_id: 4
templ_id: 5
templ_id: 6
templ_id: 7
templ_id: 8
templ_id: 9
...

• noise_test() testからtrainに変更

結果

プログラム中身

angle_test()内のtestモード内部

       std::cout << "train end" << std::endl << std::endl;
    }else if(mode=="test"){
        std::vector<std::string> ids;
        ids.push_back("test");
        detector.readClasses(ids, prefix+"case1/%s_templ.yaml");

        // angle & scale are saved here, fetched by match id
        auto infos = shape_based_matching::shapeInfo_producer::load_infos(prefix + "case1/test_info.yaml");

        Mat test_img = imread(prefix+"case1/test.png");
        assert(!test_img.empty() && "check your img path")

yamlファイルを呼び出している. => template形状などyamlファイルに保存してあるみたい

• info

• templ

以下で形状マッチを実施しているみたい

        auto matches = detector.match(img, 90, ids);

このdetectorクラスは以下のように生成され処理されている.テンプレート形状を内部に持っている.

    line2Dup::Detector detector(128, {4, 8});
      detector.readClasses(ids, prefix+"case1/%s_templ.yaml");

• 画像を変えてみる

angle_test()の画像を変更

train.pngを

test.pngを

プログラムを以下のように変更

trainモードで形状データセット作成 => testモードでマッチング処理実行

結果

探索できた...
探索にかかった時間は以下のように,z全体で0.2sくらい,すごい...

construct response map
elasped time:0.0653973s

templ match
elasped time:0.0182259s


elasped time:0.0861984s
matches.size(): 41

match.template_id: 0
match.similarity: 100

test.pngを変えてみた
結果

いい感じ...

ちなみにtrain画像作成のためにクロップするのは,webのpng画像クリップツールを使用: https://www.iloveimg.com/crop-image/crop-png

sudo docker run -it --rm -p 6080:80 ct2034/vnc-ros-kinetic-full/shape-based-matching:0.0.1

test , train分離したらできた!

docker

タブ補間効かない

https://engineer-ninaritai.com/linux-complement/

sudo apt-get install bash-completion

複数認識

top5の値を変えてあげる

次やること

• rostopicで認識位置・回転角度を出力
• 実際のネジなどで試してみる

hostpc-docker分離

アルゴリズムのみdockerに分離,.bashrcないでhostpcとdocker間の通信が必要: (参照) https://qiita.com/hoshianaaa/items/300bf653c33fa3f0ea4a のdockerとhost pc間

• hostpc

cd ~/shape_based_matching
bash start.sh

• docker

cd ~/shape_based_matching
bash start2.sh

docker 自動起動

docker_run.sh.desktopという名前のファイルを作成~/.config/autostartに置く

[Desktop Entry]
Type=Application
Exec=gnome-terminal -- bash -c "docker run -it --rm -p 6080:80 ct2034/vnc-ros-kinetic-full; exec bash"
Hidden=false
NoDisplay=false
X-GNOME-Autostart-enabled=true
Name[en_US]=start-slam
Name=start-slam
Comment[en_US]=start slam application.
Comment=start slam application.

docker image作成

FROM ros:kinetic
# install ros tutorials packages
RUN apt-get update
RUN apt-get install -y ros-kinetic-cv-bridge \
    ros-kinetic-image-transport \
    && rm -rf /var/lib/apt/lists/
RUN git clone https://github.com/hoshianaaa/shape_based_matching.git
RUN cd shape_based_matching \
   &&  mkdir build \
   &&  cd build \
   &&  cmake .. \
   &&  make

• docker file build

docker image build -t shape_based_matching .

docker image 実行 && shape_based_matching実行

docker run -it --rm shape_based_matching bash shape_based_matching/run

    1  sudo apt update
    2  cd
    3  sudo apt install ros-kinetic-cv-bridge
    4  sudo apt install ros-kinetic-cv-bridge -y
    5  his
    6  history 
    7  sudo apt install git -y
    8  git clone https://github.com/hoshianaaa/shape_based_matching.git
    9  ls
   10  cd shape_based_matching/
   11  mkdir build 
   12  cd build/
   13  cmake ..
   14  sudo apt install ros-kinetic-image-transport
   15  make
   16  ls
   17  cmake ..
   18  make
   19  history 
   20  ls
   21  ./shape_based_matching_test 
   22  cd ..
   23  source in_bashrc.sh 
   24  ./build/shape_based_matching_test 
   25  ls
   26  clear
   27  history