Merge remote-tracking branch 'upstream/main' into main

README.md

@@ -291,7 +291,11 @@ rqt, joystick, 各種コントローラーで車両操作し、学習用のデ
 
 ```
 cd ~/catkin_ws/src/ai_race/ai_race/utility/scripts
-python keyboard_con_pygame2.py
+python keyboard_con_pygame_videosave.py
+
+### `ESC`キー押下で終了
+### 終了後、学習データ（画像とコマンド）を格納したディレクトリがあることをコマンドから確認する
+ls ${HOME}/Images_from_rosbag/
 ```
 
 上記実行後、左下の「？」のうち`keyboard_con....py`が表示されてるものを押して、<br>
@@ -304,26 +308,9 @@ a 左にまがる
 d 右にまがる
 ```
 
-車両が動いている際の、画像とコマンド操作ログを取得するには以下を別ターミナルで実行します。<br>
-デフォルトでは`${HOME}`に画像とコマンド操作ログを含む`rosbag(.bag)`ファイルが出力されます。<br>
-
-```
-roslaunch sim_environment rosbag.launch
-
-### `Ctl+C`で終了する
-### 終了後、rosbag(.bag)ファイルがあることをコマンドから確認する
-ls ${HOME}/*.bag
-```
-
-`rosbag(.bag)`ファイルを、画像とコマンドに変換するには以下を実行します。<br>
-デフォルトでは`${HOME}/Images_from_rosbag/.`以下にファイルが出力されます。<br>
-
-```
-cd ~/catkin_ws/src/ai_race/ai_race/utility/script
-python rosbag_to_images_and_commands.py xxx.bag  # xxx.bagファイルは実在するrosbagファイルを指定する
-ls ${HOME}/Images_from_rosbag/.                  # 画像とコマンドの変換データがあること確認する
-```
-
+`keyboard_con_pygame_videosave.py`の停止は`ESC`キーを押下して下さい。<br>
+その後、`${HOME}/Images_from_rosbag/`以下に学習データ（画像とコマンド）が格納されます。<br>
+<br>
 以上で作成したデータを、学習モデル作成に使用下さい。
 
 ### 3.2. 各種コマンドの説明

ai_race/learning/launch/inference_from_image.launch

@@ -2,9 +2,15 @@
 <launch>
   <!-- roslaunch arguments -->
   <arg name="trt_model" default="default_path" />
+  <arg name="image_topic_name" default="/front_camera/image_exp" />
 
+  <!-- republish raw image -->
+  <include file="$(find sim_environment)/launch/image_republish.launch"/>
+
+  <!-- call node -->
   <node pkg="learning" type="rosnode_inference_from_image.py" name="rosnode_inference_from_image" output="screen" >
     <param name="trt_model" value="$(arg trt_model)" />
+    <param name="image_topic_name" value="$(arg image_topic_name)" />
   </node>
 
 </launch>

ai_race/learning/scripts/MyDataSet.py

@@ -9,9 +9,8 @@
 IMG_IDX = 1
 
 class MyDataset(Dataset):
-   def __init__(self, csv_file_path, root_dir, transform=None):
-      self.image_dataframe = pd.read_csv(csv_file_path,engine='python')
-      self.root_dir = root_dir
+   def __init__(self,image_dataframe, transform=None):
+      self.image_dataframe = image_dataframe
       self.transform = transform
 
    def __len__(self):

ai_race/learning/scripts/inference_from_image.py

@@ -121,7 +121,8 @@ def inference_from_image():
     global twist_pub
     rospy.init_node('inference_from_image', anonymous=True)
     twist_pub = rospy.Publisher('/cmd_vel', Twist, queue_size=1)
-    rospy.Subscriber("/front_camera/image_raw", Image, set_throttle_steer)
+    image_topic_name = args.image_topic_name
+    rospy.Subscriber(image_topic_name, Image, set_throttle_steer)
     r = rospy.Rate(10)
     #while not rospy.is_shutdown():
     #    r.sleep()
@@ -137,6 +138,7 @@ def parse_args():
 	arg_parser.add_argument("--trt_module", action='store_true')
 	arg_parser.add_argument("--pretrained_model", type=str, default='/home/shiozaki/work/experiments/models/checkpoints/sim_race_joycon_ResNet18_6_epoch=20.pth')
 	arg_parser.add_argument("--trt_model", type=str, default='road_following_model_trt.pth' )
+	arg_parser.add_argument("--image_topic_name", type=str, default='/front_camera/image_raw' )
 
 	args = arg_parser.parse_args()
 

ai_race/learning/scripts/rosnode_inference_from_image.py

@@ -12,10 +12,12 @@ def main():
     rospack = rospkg.RosPack()
     learning_pkg_path = rospack.get_path('learning')
     trt_model_path = rospy.get_param('~trt_model')
+    image_topic_name = rospy.get_param('~image_topic_name')
+    #image_topic_name = "/front_camera/image_exp"
 
     # call command
     inference_from_image_path = learning_pkg_path + '/scripts/inference_from_image.py'
-    command = 'python' + ' ' + inference_from_image_path + ' ' + '--trt_module --trt_model' + ' ' + trt_model_path
+    command = 'python' + ' ' + inference_from_image_path + ' ' + '--trt_module --trt_model' + ' ' + trt_model_path + ' ' + "--image_topic_name" +' ' + image_topic_name
     print(command)
     subprocess.call(command.split())
 

ai_race/learning/scripts/train.py

@@ -2,7 +2,6 @@
 import torch
 import torch.nn as nn
 import torch.optim as optim
-import torch.nn.functional as F
 import torch.backends.cudnn as cudnn
 import torchvision
 import torchvision.transforms as transforms
@@ -11,11 +10,11 @@
 import os
 import io
 import argparse
+import numpy as np
 import pandas as pd
 from sklearn.metrics import confusion_matrix
 from sklearn.metrics import classification_report
 from sklearn.metrics import f1_score
-from sklearn.model_selection import train_test_split
 from MyDataSet import MyDataset
 from samplenet import SampleNet, SimpleNet
 
@@ -26,13 +25,18 @@ def main():
 
 	# Set device.
 	device = 'cuda' if torch.cuda.is_available() else 'cpu'
-	ROOT_DIR = ""
-	imgDataset = MyDataset(args.data_csv, ROOT_DIR, transform=transforms.ToTensor())
-	# Load dataset.
-	train_data, test_data = train_test_split(imgDataset, test_size=0.2)
-	pd.to_pickle(test_data, "test_data.pkl")
-	del test_data
+
+	# Prepare dataset.
+	np.random.seed(seed=0)
+	image_dataframe = pd.read_csv(args.data_csv, engine='python', header=None)
+	image_dataframe = image_dataframe.reindex(np.random.permutation(image_dataframe.index))
+	test_num = int(len(image_dataframe) * 0.2)
+	train_dataframe = image_dataframe[test_num:]
+	test_dataframe = image_dataframe[:test_num]
+	train_data = MyDataset(train_dataframe, transform=transforms.ToTensor())
+	test_data = MyDataset(test_dataframe, transform=transforms.ToTensor())
 	train_loader = torch.utils.data.DataLoader(train_data, batch_size=20, shuffle=True)
+	test_loader = torch.utils.data.DataLoader(test_data, batch_size=20)
 
 	print('data set')
 	# Set a model.
@@ -63,19 +67,11 @@ def main():
 
 		# Output score.
 		if(epoch%args.test_interval == 0):
-			pd.to_pickle(train_data, "train_data.pkl")
-			del train_data
-
-			test_data = pd.read_pickle("test_data.pkl")
-			test_loader = torch.utils.data.DataLoader(test_data, batch_size=20, shuffle=True)
-			del test_data
 			test_acc, test_loss = test(model, device, test_loader, criterion)
-			del test_loader
 
 			stdout_temp = 'epoch: {:>3}, train acc: {:<8}, train loss: {:<8}, test acc: {:<8}, test loss: {:<8}'
 			print(stdout_temp.format(epoch+1, train_acc, train_loss, test_acc, test_loss))
 
-			train_data = pd.read_pickle("train_data.pkl")
 		else:	
 			stdout_temp = 'epoch: {:>3}, train acc: {:<8}, train loss: {:<8}' #, test acc: {:<8}, test loss: {:<8}'
 			print(stdout_temp.format(epoch+1, train_acc, train_loss)) #, test_acc, test_loss))

ai_race/sim_environment/launch/image_republish.launch

@@ -0,0 +1,26 @@
+<?xml version="1.0"?>
+<launch>
+  <arg name="input_image_topic" default="/front_camera/image_raw" />
+  <arg name="output_image_topic" default="/front_camera/image_exp" />
+  <arg name="use_compressed" default="true" />
+  <arg name="use_image_view" default="false" />
+
+  <group if="$(arg use_compressed)">
+    <node name="image_republish" pkg="image_transport" type="republish" args="compressed raw">
+      <remap from="in" to="$(arg input_image_topic)" />
+      <remap from="out" to="$(arg output_image_topic)" />
+    </node>
+    <group if="$(eval use_image_view==true)"> 
+      <node name="image_view" pkg="image_view" type="image_view" >
+	<remap from="image" to="$(arg output_image_topic)"/>
+      </node>
+    </group>
+  </group>
+  <group unless="$(arg use_compressed)">
+    <group if="$(eval use_image_view==true)"> 
+      <node name="image_view" pkg="image_view" type="image_view" >
+	<remap from="image" to="$(arg input_image_topic)"/>
+      </node>
+    </group>
+  </group>
+</launch>

ai_race/sim_environment/scripts/judge_communication_d.py

@@ -19,13 +19,13 @@ class JudgeCommunicationd(object):
     #jsonの内容
         # judge_info
 
-    def __init__(self, judge_url, GameStateCallback_timer_duration):
+    def __init__(self, judge_url, GameStateCallback_timer_duration, Post_Ros_timer_duration):
         self.judge_url = judge_url
         self.vel_pub = rospy.Publisher('gamestate', String, queue_size=1)
         #print(GameStateCallback_timer_duration)
         rospy.Timer(rospy.Duration(GameStateCallback_timer_duration),
                     self.publishGameStateCallback)
-        rospy.Timer(rospy.Duration(GameStateCallback_timer_duration),
+        rospy.Timer(rospy.Duration(Post_Ros_timer_duration),
                     self.postRosTimerCallback)
         self.tryCourseRecoveryCallback = rospy.Subscriber('gamestate', String, self.tryCourseRecoveryCallback)
 
@@ -98,7 +98,8 @@ def tryCourseRecoveryCallback(self, state):
     JUDGE_URL = rospy.get_param('~judge_url', 'http://127.0.0.1:5000')
     JUDGESERVER_GETSTATE_URL = JUDGE_URL + "/judgeserver/getState"
     TIMER_DURATION = rospy.get_param('~GameStateCallback_timer_duration', 0.25)
-    JUDGECOMMUNICATIOND = JudgeCommunicationd(JUDGESERVER_GETSTATE_URL, TIMER_DURATION)
+    POST_ROSTIMER_DURATION = rospy.get_param('~PostRosTimerCallback_timer_duration', 0.25)
+    JUDGECOMMUNICATIOND = JudgeCommunicationd(JUDGESERVER_GETSTATE_URL, TIMER_DURATION, POST_ROSTIMER_DURATION)
 
     rospy.spin()