when detection is none i got this error: object of type 'NoneType' has no len()

[Rasantis]

Search before asking

• I have searched the Supervision issues and found no similar feature requests.

Question

I´m running this code, when the yolo detects something on the video its fine, it works, but when there is no detections, i got this error:
object of type 'NoneType' has no len()
and the code stop working, in this code i was expectating that him keep working, even if there´s no detection..... this is my code:

Additional

import cv2
import supervision as sv
import json
from ultralytics import YOLO
import os
import paho.mqtt.client as mqtt
import numpy as np
 
class LineCounter:
    def __init__(self, lines_json_path, model_path, video_source, original_width, original_height):
        with open(lines_json_path, 'r') as f:
            self.lines_data = json.load(f)
        self.model = YOLO(model_path)
        self.video_source = video_source
        self.original_width = original_width
        self.original_height = original_height
        self.last_sent_data = {} 
        self.mqtt_client = mqtt.Client("UniqueClientName") 
        self.mqtt_client.connect("localhost", 1883)  

        self.initialize_counters()

        self.line_annotator = sv.LineZoneAnnotator(
            thickness=2, 
            text_thickness=2, 
            text_scale=3, 
            custom_in_text="entrando",
            custom_out_text="saindo"
        )
        
        self.box_annotator = sv.BoxAnnotator(
            thickness=2,
            text_thickness=1,
            text_scale=0.5
        )

    def initialize_counters(self):
        self.line_counters = {}
        for result in self.model.track(source=self.video_source, show=False, stream=True, agnostic_nms=True):
            frame = result.orig_img
            self.current_width, self.current_height = frame.shape[1], frame.shape[0]
            break

        self.width_scale = self.current_width / self.original_width
        self.height_scale = self.current_height / self.original_height

        for line_key, line_value in self.lines_data.items():
            adjusted_height_scale = self.height_scale * 2.37
            start_point_x, start_point_y = line_value['points'][0]
            end_point_x, end_point_y = line_value['points'][1]
            start_point = sv.Point(int(start_point_x * self.width_scale * 2), int(start_point_y * adjusted_height_scale))
            end_point = sv.Point(int(end_point_x * self.width_scale * 2), int(end_point_y * adjusted_height_scale))
            self.line_counters[line_key] = sv.LineZone(start=start_point, end=end_point)

    def resize_frame(self, frame):
        max_width = 600
        max_height = 400
        aspect_ratio = frame.shape[1] / frame.shape[0]
        
        if aspect_ratio > 1:
            new_width = max_width
            new_height = int(max_width / aspect_ratio)
        else:
            new_height = max_height
            new_width = int(max_height * aspect_ratio)

        return cv2.resize(frame, (new_width, new_height))

    def run(self, output_directory, output_filename):
        os.makedirs(output_directory, exist_ok=True)
        fourcc = cv2.VideoWriter_fourcc("m", "p", "4", "v")
        first_frame = True
        out = None

        for result in self.model.track(source=self.video_source, show=False, stream=True, agnostic_nms=True):
            frame = result.orig_img
            detections = sv.Detections.from_ultralytics(result)
            detections = detections[detections.class_id == 0] # ABENÇOADA LINHA RESPONSÁVEL POR FAZER A FILTRAGEM DE QUAL CLASSE DETECTAR

            for line_key in self.line_counters.keys():
                self.line_counters[line_key].trigger(detections=detections)
                in_count = self.line_counters[line_key].in_count
                out_count = self.line_counters[line_key].out_count

                # Criando o payload
                payload = {
                    "line": line_key,
                    "in_count": in_count,
                    "out_count": out_count
                }

                # Verifica se os dados são diferentes dos últimos dados enviados
                last_data = self.last_sent_data.get(line_key, {})
                if payload != last_data:
                    # Enviando dados para o MQTT
                    self.mqtt_client.publish("sensor/linha", json.dumps(payload))
                    # Atualiza o último valor enviado
                    self.last_sent_data[line_key] = payload
            
            if first_frame:
                resized_frame = self.resize_frame(frame)
                out = cv2.VideoWriter(f"{output_directory}/{output_filename}", fourcc, 30.0, (resized_frame.shape[1], resized_frame.shape[0]))
                if out.isOpened():
                    print("VideoWriter initialized successfully.")
                else:
                    print("Failed to initialize VideoWriter.")
                    return
                first_frame = False

            labels = [
                f"{detections.tracker_id[i]} {self.model.model.names[detections.class_id[i]]} {detections.confidence[i]:0.2f}"
                for i in range(len(detections.tracker_id))
            ]
            
            frame = self.box_annotator.annotate(
                scene=frame, 
                detections=detections,
                labels=labels
            )

            total_count = 0
            for line_key in self.line_counters.keys():
                self.line_annotator.annotate(frame=frame, line_counter=self.line_counters[line_key])
                total_count += self.line_counters[line_key].in_count + self.line_counters[line_key].out_count

            resized_frame = self.resize_frame(frame)
            out.write(resized_frame)
            cv2.imshow("yolov8", resized_frame)

            if cv2.waitKey(30) == ord('q'):
                break

        out.release()
        cv2.destroyAllWindows()

[SkalskiP]

Hi, @Rasantis 👋🏻 let' me convert that question into a discussion and move it to our Q&A section.

[SkalskiP]

@Rasantis It will be pretty challenging for us to help you out unless you provide us with minimal reproducible examples, a full error stack trace, and a video file that you used.