Disable detection

requirements.txt

@@ -1,6 +1,6 @@
-scipy==1.6.3
-filterpy==1.4.5
-pygame==2.0.1
-pytest==6.2.3
-PyYAML==5.4.1
-transitions==0.8.10
+scipy
+filterpy
+pygame
+pytest
+PyYAML
+transitions

src/copilot.py

@@ -3,10 +3,10 @@
 
 from PIL import Image
 
-from pycoral.adapters import common
-from pycoral.adapters import detect
-from pycoral.adapters import classify
-from pycoral.utils.dataset import read_label_file
+# from pycoral.adapters import common
+# from pycoral.adapters import detect
+# from pycoral.adapters import classify
+# from pycoral.utils.dataset import read_label_file
 
 # from .button import Button
 from .utils import (
@@ -56,7 +56,7 @@ def __init__(
 
  self._classfication_interpreter = traffic_light_classifier_interpreter
  self._classfication_interpreter.allocate_tensors()
- self._traffic_light_size = common.input_size(self._classfication_interpreter)
+ # self._traffic_light_size = common.input_size(self._classfication_interpreter)
 
  self._speaker = speaker
 
@@ -66,12 +66,12 @@ def __init__(
  assert tile_size == input_shape[1]
  self._tile_config = TileConfig(tile_size, tile_w_overlap, tile_h_overlap)
 
- self._ssd_labels = read_label_file(self._args.label) if self._args.label else {}
- self._traffic_light_labels = (
- read_label_file(self._args.traffic_light_label)
- if self._args.traffic_light_label
- else {}
- )
+ # self._ssd_labels = read_label_file(self._args.label) if self._args.label else {}
+ # self._traffic_light_labels = (
+ # read_label_file(self._args.traffic_light_label)
+ # if self._args.traffic_light_label
+ # else {}
+ # )
 
  self._h_crop_keep_percentage = 0.6
  self._led = led
@@ -97,7 +97,7 @@ def run(self):
  )
  prev_cycle_time = current_cycle_time
  logging.debug("recv image from: {}".format(image_time))
- self.process(image)
+ # self.process(image)
  logging.debug(
  "process time: %.2f ms"
  % ((time.perf_counter() - current_cycle_time) * 1000)
@@ -142,7 +142,7 @@ def classify(self, traffic_light_thumbnail):
  traffic_light_resized = traffic_light_thumbnail.resize(
  self._traffic_light_size, Image.ANTIALIAS
  )
- common.set_input(self._classfication_interpreter, traffic_light_resized)
+ # common.set_input(self._classfication_interpreter, traffic_light_resized)
  start = time.perf_counter()
  self._classfication_interpreter.invoke()
  classes = classify.get_classes(
@@ -179,7 +179,7 @@ def detect(self, img):
  ):
  # print(tile_location)
  tile = img.crop(tile_location)
- common.set_input(self._ssd_interpreter, tile)
+ # common.set_input(self._ssd_interpreter, tile)
  start = time.perf_counter()
  self._ssd_interpreter.invoke()
  inference_time += time.perf_counter() - start

src/main.py

@@ -4,10 +4,9 @@
 import time
 import threading
 
-import picamera
+from picamera2 import Picamera2
 
 from src.os_utils import generate_recording_postfix
-from .led import Led
 from .camera_capturer import CameraCapturer
 from .camera_recorder import CameraRecorder
 from .pubsub import PubSub
@@ -109,40 +108,45 @@ def main():
 
  blackbox = BlackBox(image_saver)
 
- with picamera.PiCamera() as camera:
- # fps for recording
- camera.framerate = 20
- camera.resolution = camera_info.resolution
- camera.exposure_mode = "sports"
-
- led_pin = 10
- led = Led(led_pin)
- camera_recorder = CameraRecorder(camera, led, args.blackbox_path)
- camera_capturer = CameraCapturer(
- camera, 5, camera_recorder.is_recording, pubsub, inference_config
- )
+ camera = Picamera2()
+
+
+ main_stream = {"size": camera_info.resolution}
+ lores_stream = {"size": inference_config.inference_resolution}
+ video_config = camera.create_video_configuration(main_stream, lores_stream, encode="main")
+ camera.configure(video_config)
+
+ #camera.framerate = 20
+ #camera.exposure_mode = "sports"
 
- if args.cpu:
- from tflite_runtime.interpreter import Interpreter as make_interpreter
- else:
- from pycoral.utils.edgetpu import make_interpreter
-
- try:
- copilot = CoPilot(
- args,
- pubsub,
- blackbox,
- camera_info,
- inference_config,
- led,
- Speaker(args.lang),
- make_interpreter(args.ssd_model),
- make_interpreter(args.traffic_light_classification_model),
- )
- except ValueError as e:
- print(str(e) + "Use --cpu if you do not have a coral tpu")
- return
- copilot.run()
+ led_pin = 10
+ led = ILed()
+ camera_recorder = CameraRecorder(camera, led, args.blackbox_path)
+ camera_capturer = CameraCapturer(
+ camera, 5, camera_recorder.is_recording, pubsub, inference_config
+ )
+
+ if args.cpu:
+ from tflite_runtime.interpreter import Interpreter as make_interpreter
+ else:
+ from pycoral.utils.edgetpu import make_interpreter
+
+ try:
+ copilot = CoPilot(
+ args,
+ pubsub,
+ blackbox,
+ camera_info,
+ inference_config,
+ led,
+ Speaker(args.lang),
+ make_interpreter(args.ssd_model),
+ make_interpreter(args.traffic_light_classification_model),
+ )
+ except ValueError as e:
+ print(str(e) + "Use --cpu if you do not have a coral tpu")
+ return
+ copilot.run()
 
  except Exception as e:
  Speaker(args.lang).play_sound("dead", is_blocking=True)