Number of rectangles drawn using opencv in request_callback

[krockamichael]

Hello, I want to ask how many rectangles can be drawn using a custom DrawRectangles request_callback - like the one in examples/tensorflow/real_time.py. I have a picamera2 set up like this, so that I can show the camera "view" (output?) in my flask web application:

from picamera2.encoders.jpeg_encoder import JpegEncoder
from picamera2.picamera2 import *
from threading import Condition
import io

class StreamingOutput(io.BufferedIOBase):
    def __init__(self):
        self.frame = None
        self.condition = Condition()
    def write(self, buf):
        with self.condition:
            self.frame = buf
            self.condition.notify_all()


def gather_img( output):
    while True:
        with output.condition:
            # time.sleep(0.1)
            output.condition.wait()
            frame = output.frame
        yield (b'--frame\r\nContent-Type:image/jpeg\r\n\r\n' + frame + b'\r\n')

def DrawRectangles(request):
    global points, colors, lp_dict  # defined elsewhere
    stream = request.picam2.stream_map['main']
    fb = request.request.buffers[stream]

    with fb.mmap(0) as b:
        im = np.array(b, copy=False, dtype=np.uint8).reshape((480, 640, 4)) # (960, 1280) (1440, 1920)

        for i, value in enumerate(zip(colors, points)):
            color, point = value[0], value[1]
            cv2.rectangle(im, point[0], point[1], color)

            # coords are top center of bounding box for parking space
            coords = (int(point[0][0] + (point[1][0] - point[0][0]) / 2), point[0][1])
            cv2.putText(im, str(i), coords, cv2.FONT_HERSHEY_SIMPLEX, 1, (255,255,255), 2, cv2.LINE_AA)

            if i in lp_dict:
                coords = (point[0][0], point[0][1] + (point[0][1] - point[1][1]))
                cv2.putText(im, lp_dict[i], coords, cv2.FONT_HERSHEY_SIMPLEX, 1, (255,255,255), 2, cv2.LINE_AA)

        del im

# "main"
if picam2 is None:
        picam2 = Picamera2()
        picam2.start_preview()
        config = picam2.video_configuration(main={"size": (640, 480)},  # (1280, 960) (1920, 1440)
                                            lores={"size": (320, 240), "format": "YUV420"})
        picam2.request_callback = DrawRectangles
        picam2.configure(config)
    output = StreamingOutput()
    picam2.start_recording(JpegEncoder(), output)

    return Response(gather_img(output), mimetype='multipart/x-mixed-replace; boundary=frame')

The number of rectangles I attempt to display is 12, but with this setup I am only able to show 3 consistently while 3 more flicker, and the remaining 6 are never displayed. This is my HTML template.

{% extends "base.html" %}

{% block content %}
<h1 class="title">
  Parking Space Occupancy Detector
</h1>

<h2 class="subtitle">
  Raspberry Pi 4 - HQ camera - Wide-angle lens
</h2>

{% with messages = get_flashed_messages() %}
{% if messages %}
<div class="notification is-danger" style="width:30%; margin: auto;">
  {{ messages[0] }}
</div>
{% endif %}
{% endwith %}

<div>
  <img src={{ url_for('main.camera')  }} width="640" height="480">
</div>

<a class="button" href="{{ url_for('main.predict') }}">Tflite model & email test!</a>

{% endblock %}

[krockamichael]

Some of the effects can be alleviated by increasing the resolution - in particular width="2240" height="1680" seems to work "best" - as in the rectangles do not flicker (mostly, about once per 10 seconds). This comes at the cost of "laggy" mjpeg stream, which in my particular case is not as big a problem as non-visible / flickering rectangles.