feat(autoware_traffic_light_classifier): add traffic light classifier schema, README and car and ped launcher

launch/tier4_perception_launch/launch/traffic_light_recognition/traffic_light_node_container.launch.py

@@ -97,7 +97,7 @@ def create_parameter_dict(*args):
                 plugin="autoware::traffic_light::TrafficLightClassifierNodelet",
                 name="car_traffic_light_classifier",
                 namespace="classification",
-                parameters=[car_traffic_light_classifier_model_param],
+                parameters=[car_traffic_light_classifier_model_param, {"build_only": False}],
                 remappings=[
                     ("~/input/image", camera_arguments["input/image"]),
                     ("~/input/rois", camera_arguments["output/rois"]),
@@ -112,7 +112,7 @@ def create_parameter_dict(*args):
                 plugin="autoware::traffic_light::TrafficLightClassifierNodelet",
                 name="pedestrian_traffic_light_classifier",
                 namespace="classification",
-                parameters=[pedestrian_traffic_light_classifier_model_param],
+                parameters=[pedestrian_traffic_light_classifier_model_param, {"build_only": False}],
                 remappings=[
                     ("~/input/image", camera_arguments["input/image"]),
                     ("~/input/rois", camera_arguments["output/rois"]),
@@ -179,7 +179,7 @@ def create_parameter_dict(*args):
                 plugin="autoware::traffic_light::TrafficLightFineDetectorNode",
                 name="traffic_light_fine_detector",
                 namespace=f"{namespace}/detection",
-                parameters=[fine_detector_model_param],
+                parameters=[fine_detector_model_param, {"build_only": False}],
                 remappings=[
                     ("~/input/image", camera_arguments["input/image"]),
                     ("~/input/rois", "rough/rois"),
@@ -227,16 +227,14 @@ def add_launch_arg(name: str, default_value=None, description=None):
     # traffic_light_classifier
     add_launch_arg(
         "car_classifier_param_path",
-        os.path.join(
-            classifier_share_dir, "config", "car_traffic_light_classifier_efficientNet.param.yaml"
-        ),
+        os.path.join(classifier_share_dir, "config", "car_traffic_light_classifier.param.yaml"),
     )
     add_launch_arg(
         "pedestrian_classifier_param_path",
         os.path.join(
             classifier_share_dir,
             "config",
-            "pedestrian_traffic_light_classifier_efficientNet.param.yaml",
+            "pedestrian_traffic_light_classifier.param.yaml",
         ),
     )
 

perception/autoware_traffic_light_classifier/README.md

@@ -63,25 +63,19 @@ These colors and shapes are assigned to the message as follows:
 
 ### Node Parameters
 
-| Name                          | Type   | Description                                                                                                                                                                                                                                                                                                                        |
-| ----------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `classifier_type`             | int    | If the value is `1`, cnn_classifier is used                                                                                                                                                                                                                                                                                        |
-| `data_path`                   | str    | Packages data and artifacts directory path                                                                                                                                                                                                                                                                                         |
-| `backlight_threshold`         | double | If the intensity of light is grater than this threshold, the color and shape of the corresponding ROI will be overwritten with UNKNOWN, and the confidence of the overwritten signal will be set to `0.0`. The value should be set in the range of `[0.0, 1.0]`. If you wouldn't like to use this feature, please set it to `1.0`. |
-| `classify_traffic_light_type` | int    | If the value is `0`, vehicular signals are classified. If the value is `1`, pedestrian signals are classified.                                                                                                                                                                                                                     |
+#### car_traffic_light_classifier
+
+{{ json_to_markdown("perception/autoware_traffic_light_classifier/schema/car_traffic_light_classifier.schema.json") }}
+
+#### pedestrian_traffic_light_classifier
+
+{{ json_to_markdown("perception/autoware_traffic_light_classifier/schema/pedestrian_traffic_light_classifier.schema.json") }}
 
 ### Core Parameters
 
 #### cnn_classifier
 
-| Name                    | Type            | Description                          |
-| ----------------------- | --------------- | ------------------------------------ |
-| `classifier_label_path` | str             | path to the model file               |
-| `classifier_model_path` | str             | path to the label file               |
-| `classifier_precision`  | str             | TensorRT precision, `fp16` or `int8` |
-| `classifier_mean`       | vector\<double> | 3-channel input image mean           |
-| `classifier_std`        | vector\<double> | 3-channel input image std            |
-| `apply_softmax`         | bool            | whether or not apply softmax         |
+Including [this section](#car_traffic_light_classifier)
 
 #### hsv_classifier
 

perception/autoware_traffic_light_classifier/config/car_traffic_light_classifier.param.yaml

@@ -7,5 +7,5 @@
     classifier_mean: [123.675, 116.28, 103.53]
     classifier_std: [58.395, 57.12, 57.375]
     backlight_threshold: 0.85
-    classifier_type: 1 #classifier_type {hsv_filter: 0, cnn: 1}
-    classify_traffic_light_type: 0 #classify_traffic_light_type {car: 0, pedestrian:1}
+    classifier_type: 1
+    classify_traffic_light_type: 0

perception/autoware_traffic_light_classifier/config/pedestrian_traffic_light_classifier.param.yaml

@@ -7,5 +7,5 @@
     classifier_mean: [123.675, 116.28, 103.53]
     classifier_std: [58.395, 57.12, 57.375]
     backlight_threshold: 0.85
-    classifier_type: 1 #classifier_type {hsv_filter: 0, cnn: 1}
-    classify_traffic_light_type: 1 #classify_traffic_light_type {car: 0, pedestrian:1}
+    classifier_type: 1
+    classify_traffic_light_type: 1

perception/autoware_traffic_light_classifier/launch/traffic_light_classifier.launch.xml → perception/autoware_traffic_light_classifier/launch/car_traffic_light_classifier.launch.xml

@@ -2,7 +2,11 @@
   <arg name="input/image" default="~/image_raw"/>
   <arg name="input/rois" default="~/rois"/>
   <arg name="output/traffic_signals" default="classified/traffic_signals"/>
-  <arg name="traffic_light_classifier_param_path" default="$(find-pkg-share autoware_traffic_light_classifier)/config/traffic_light_classifier.param.yaml" description="classifier param path"/>
+  <arg name="data_path" default="$(env HOME)/autoware_data" description="packages data and artifacts directory path"/>
+  <arg name="traffic_light_classifier_param_path" default="$(find-pkg-share autoware_traffic_light_classifier)/config/car_traffic_light_classifier.param.yaml" description="classifier param path"/>
+  <arg name="traffic_light_classifier_model_path" default="$(var data_path)/traffic_light_classifier" description="path to car onnx model and label"/>
+  <arg name="car_traffic_light_classifier_label_name" default="lamp_labels.txt" description="car label name"/>
+  <arg name="car_traffic_light_classifier_model_name" default="traffic_light_classifier_mobilenetv2_batch_6" description="car onnx model name"/>
   <arg name="build_only" default="false" description="exit after trt engine is built"/>
 
   <node pkg="autoware_traffic_light_classifier" exec="traffic_light_classifier_node" name="traffic_light_classifier" output="screen">

perception/autoware_traffic_light_classifier/launch/pedestrian_traffic_light_classifier.launch.xml

@@ -0,0 +1,23 @@
+<launch>
+  <arg name="input/image" default="~/image_raw"/>
+  <arg name="input/rois" default="~/rois"/>
+  <arg name="output/traffic_signals" default="classified/traffic_signals"/>
+  <arg name="data_path" default="$(env HOME)/autoware_data" description="packages data and artifacts directory path"/>
+  <arg
+    name="traffic_light_classifier_param_path"
+    default="$(find-pkg-share autoware_traffic_light_classifier)/config/pedestrian_traffic_light_classifier.param.yaml"
+    description="classifier param path"
+  />
+  <arg name="traffic_light_classifier_model_path" default="$(var data_path)/traffic_light_classifier" description="path to pedestrian onnx model and label"/>
+  <arg name="pedestrian_traffic_light_classifier_label_name" default="lamp_labels_ped.txt" description="pedestrian label name"/>
+  <arg name="pedestrian_traffic_light_classifier_model_name" default="ped_traffic_light_classifier_mobilenetv2_batch_6" description="pedestrian onnx model name"/>
+  <arg name="build_only" default="false" description="exit after trt engine is built"/>
+
+  <node pkg="autoware_traffic_light_classifier" exec="traffic_light_classifier_node" name="traffic_light_classifier" output="screen">
+    <remap from="~/input/image" to="$(var input/image)"/>
+    <remap from="~/input/rois" to="$(var input/rois)"/>
+    <remap from="~/output/traffic_signals" to="$(var output/traffic_signals)"/>
+    <param from="$(var traffic_light_classifier_param_path)" allow_substs="true"/>
+    <param name="build_only" value="$(var build_only)"/>
+  </node>
+</launch>

perception/autoware_traffic_light_classifier/schema/car_traffic_light_classifier.schema.json

@@ -0,0 +1,89 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "title": "autoware_traffic_light_classifier paramter",
+  "type": "object",
+  "definitions": {
+    "car_traffic_light_classifier": {
+      "type": "object",
+      "properties": {
+        "approximate_sync": {
+          "type": "boolean",
+          "description": "Enable or disable approximate synchronization.",
+          "default": false
+        },
+        "classifier_label_path": {
+          "type": "string",
+          "description": "Path to the label file for the traffic light classifier.",
+          "default": "$(var traffic_light_classifier_model_path)/$(var car_traffic_light_classifier_label_name)"
+        },
+        "classifier_model_path": {
+          "type": "string",
+          "description": "Path to the ONNX model file for the traffic light classifier.",
+          "default": "$(var traffic_light_classifier_model_path)/$(var car_traffic_light_classifier_model_name).onnx"
+        },
+        "classifier_precision": {
+          "type": "string",
+          "description": "Precision used for traffic light classifier inference. Valid values: [fp32, fp16, int8].",
+          "default": "fp16"
+        },
+        "classifier_mean": {
+          "type": "array",
+          "description": "Mean values used for input normalization [R, G, B].",
+          "items": {
+            "type": "number"
+          },
+          "default": [123.675, 116.28, 103.53]
+        },
+        "classifier_std": {
+          "type": "array",
+          "description": "Standard deviation values used for input normalization [R, G, B].",
+          "items": {
+            "type": "number"
+          },
+          "default": [58.395, 57.12, 57.375]
+        },
+        "backlight_threshold": {
+          "type": "number",
+          "description": "If the intensity get grater than this overwrite with UNKNOWN in corresponding RoI. Note that, if the value is much higher, the node only overwrites in the harsher backlight situations. Therefore, If you wouldn't like to use this feature set this value to `1.0`. The value can be `[0.0, 1.0]`. The confidence of overwritten signal is set to `0.0`.",
+          "default": 0.85
+        },
+        "classifier_type": {
+          "type": "integer",
+          "description": "Type of classifier used. {0: hsv_filter, 1: cnn}.",
+          "default": 1
+        },
+        "classify_traffic_light_type": {
+          "type": "integer",
+          "description": "Type of traffic light to classify. {0: car, 1: pedestrian}.",
+          "default": 0
+        }
+      },
+      "required": [
+        "approximate_sync",
+        "classifier_label_path",
+        "classifier_model_path",
+        "classifier_precision",
+        "classifier_mean",
+        "classifier_std",
+        "backlight_threshold",
+        "classifier_type",
+        "classify_traffic_light_type"
+      ],
+      "additionalProperties": false
+    }
+  },
+  "properties": {
+    "/**": {
+      "type": "object",
+      "properties": {
+        "ros__parameters": {
+          "$ref": "#/definitions/car_traffic_light_classifier"
+        }
+      },
+      "required": ["ros__parameters"],
+      "additionalProperties": false
+    }
+  },
+  "required": ["/**"],
+  "additionalProperties": false
+}

perception/autoware_traffic_light_classifier/schema/pedestrian_traffic_light_classifier.schema.json

@@ -0,0 +1,89 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "title": "autoware_traffic_light_classifier paramter",
+  "type": "object",
+  "definitions": {
+    "pedestrian_traffic_light_classifier": {
+      "type": "object",
+      "properties": {
+        "approximate_sync": {
+          "type": "boolean",
+          "description": "Enable or disable approximate synchronization.",
+          "default": false
+        },
+        "classifier_label_path": {
+          "type": "string",
+          "description": "Path to the label file for the traffic light classifier.",
+          "default": "$(var traffic_light_classifier_model_path)/$(var pedestrian_traffic_light_classifier_label_name)"
+        },
+        "classifier_model_path": {
+          "type": "string",
+          "description": "Path to the ONNX model file for the traffic light classifier.",
+          "default": "$(var traffic_light_classifier_model_path)/$(var pedestrian_traffic_light_classifier_model_name).onnx"
+        },
+        "classifier_precision": {
+          "type": "string",
+          "description": "Precision used for traffic light classifier inference. Valid values: [fp32, fp16, int8].",
+          "default": "fp16"
+        },
+        "classifier_mean": {
+          "type": "array",
+          "description": "Mean values used for input normalization [R, G, B].",
+          "items": {
+            "type": "number"
+          },
+          "default": [123.675, 116.28, 103.53]
+        },
+        "classifier_std": {
+          "type": "array",
+          "description": "Standard deviation values used for input normalization [R, G, B].",
+          "items": {
+            "type": "number"
+          },
+          "default": [58.395, 57.12, 57.375]
+        },
+        "backlight_threshold": {
+          "type": "number",
+          "description": "If the intensity get grater than this overwrite with UNKNOWN in corresponding RoI. Note that, if the value is much higher, the node only overwrites in the harsher backlight situations. Therefore, If you wouldn't like to use this feature set this value to `1.0`. The value can be `[0.0, 1.0]`. The confidence of overwritten signal is set to `0.0`.",
+          "default": 0.85
+        },
+        "classifier_type": {
+          "type": "integer",
+          "description": "Type of classifier used. {0: hsv_filter, 1: cnn}.",
+          "default": 1
+        },
+        "classify_traffic_light_type": {
+          "type": "integer",
+          "description": "Type of traffic light to classify. {0: car, 1: pedestrian}.",
+          "default": 1
+        }
+      },
+      "required": [
+        "approximate_sync",
+        "classifier_label_path",
+        "classifier_model_path",
+        "classifier_precision",
+        "classifier_mean",
+        "classifier_std",
+        "backlight_threshold",
+        "classifier_type",
+        "classify_traffic_light_type"
+      ],
+      "additionalProperties": false
+    }
+  },
+  "properties": {
+    "/**": {
+      "type": "object",
+      "properties": {
+        "ros__parameters": {
+          "$ref": "#/definitions/pedestrian_traffic_light_classifier"
+        }
+      },
+      "required": ["ros__parameters"],
+      "additionalProperties": false
+    }
+  },
+  "required": ["/**"],
+  "additionalProperties": false
+}

perception/autoware_traffic_light_classifier/src/classifier/cnn_classifier.cpp

@@ -34,15 +34,13 @@ CNNClassifier::CNNClassifier(rclcpp::Node * node_ptr) : node_ptr_(node_ptr)
   std::string precision;
   std::string label_file_path;
   std::string model_file_path;
-  precision = node_ptr_->declare_parameter("classifier_precision", "fp16");
-  label_file_path = node_ptr_->declare_parameter("classifier_label_path", "labels.txt");
-  model_file_path = node_ptr_->declare_parameter("classifier_model_path", "model.onnx");
+  precision = node_ptr_->declare_parameter<std::string>("classifier_precision");
+  label_file_path = node_ptr_->declare_parameter<std::string>("classifier_label_path");
+  model_file_path = node_ptr_->declare_parameter<std::string>("classifier_model_path");
   // ros param does not support loading std::vector<float>
   // we have to load std::vector<double> and transfer to std::vector<float>
-  auto mean_d =
-    node_ptr->declare_parameter("classifier_mean", std::vector<double>{123.675, 116.28, 103.53});
-  auto std_d =
-    node_ptr->declare_parameter("classifier_std", std::vector<double>{58.395, 57.12, 57.375});
+  auto mean_d = node_ptr->declare_parameter<std::vector<double>>("classifier_mean");
+  auto std_d = node_ptr->declare_parameter<std::vector<double>>("classifier_std");
   mean_ = std::vector<float>(mean_d.begin(), mean_d.end());
   std_ = std::vector<float>(std_d.begin(), std_d.end());
   if (mean_.size() != 3 || std_.size() != 3) {
@@ -55,7 +53,7 @@ CNNClassifier::CNNClassifier(rclcpp::Node * node_ptr) : node_ptr_(node_ptr)
   classifier_ = std::make_unique<autoware::tensorrt_classifier::TrtClassifier>(
     model_file_path, precision, mean_, std_);
   batch_size_ = classifier_->getBatchSize();
-  if (node_ptr_->declare_parameter("build_only", false)) {
+  if (node_ptr_->declare_parameter<bool>("build_only")) {
     RCLCPP_INFO(node_ptr_->get_logger(), "TensorRT engine is built and shutdown node.");
     rclcpp::shutdown();
   }

perception/autoware_traffic_light_classifier/src/traffic_light_classifier_node.cpp

@@ -25,11 +25,11 @@ namespace autoware::traffic_light
 TrafficLightClassifierNodelet::TrafficLightClassifierNodelet(const rclcpp::NodeOptions & options)
 : Node("traffic_light_classifier_node", options)
 {
-  classify_traffic_light_type_ = this->declare_parameter("classify_traffic_light_type", 0);
+  classify_traffic_light_type_ = this->declare_parameter<int>("classify_traffic_light_type");
 
   using std::placeholders::_1;
   using std::placeholders::_2;
-  is_approximate_sync_ = this->declare_parameter("approximate_sync", false);
+  is_approximate_sync_ = this->declare_parameter<bool>("approximate_sync");
   backlight_threshold_ = this->declare_parameter<double>("backlight_threshold");
 
   if (is_approximate_sync_) {
@@ -49,8 +49,7 @@ TrafficLightClassifierNodelet::TrafficLightClassifierNodelet(const rclcpp::NodeO
   timer_ = rclcpp::create_timer(
     this, get_clock(), 100ms, std::bind(&TrafficLightClassifierNodelet::connectCb, this));
 
-  int classifier_type = this->declare_parameter(
-    "classifier_type", static_cast<int>(TrafficLightClassifierNodelet::ClassifierType::HSVFilter));
+  int classifier_type = this->declare_parameter<int>("classifier_type");
   if (classifier_type == TrafficLightClassifierNodelet::ClassifierType::HSVFilter) {
     classifier_ptr_ = std::make_shared<ColorClassifier>(this);
   } else if (classifier_type == TrafficLightClassifierNodelet::ClassifierType::CNN) {