00:00Wait, look at the picture and tell me how many objects you can spot in 3 seconds.
00:06How many objects do you think this AI object detection setup will detect at the same time?
00:10It's hard to find all of them, right?
00:12These are the results of the image from the object detection setup.
00:14Not only this, it can detect, book, cell phone, track, while cycle and the list goes on up to 75
00:20plus classes.
00:21Pretty cool, right?
00:22Come let's build it from stats and test it.
00:24This is our setup.
00:25We use the HP 22 cam and push button.
00:28The push button is connected to the 13th pin of the HP 22 cam.
00:31Next, open your browser and go to Secure Regist Cloud.
00:34Log in, if you don't have an account, create one.
00:37Scroll down and click on object detection.
00:40Inside you will see my usage, my logs and API.
00:44And there is a class selection section where you can choose specific objects to detect or select all 75 classes.
00:50Then confidence title, which you can set according to your need.
00:53Now, let's test it without any hardware.
00:57Upload any random image in try API and click run.
01:00And instantly, you will see detected object, bounding boxes and confidence values.
01:05Now, let's bring this into real life with the real hardware.
01:09Scroll down to the code section and select ESP32 cam.
01:12There is also code for Raspberry Pi, shovel.
01:15So you can select any board and any code according to your need.
01:18Copy the code and change the Wi-Fi name and Wi-Fi password.
01:21Your API key will be auto-filled.
01:24Upload the code.
01:25Once it's powered on, the Wi-Fi is connected.
01:28Now, place the camera facing any scene.
01:30Press the button.
01:31The ESP32 cam captures an image.
01:34Send it to the API cloud.
01:36And within seconds, you can get object names, number of objects detected, confidence code for each.
01:41This is how small meets smart.
01:44Check out the comment section and bio for full tutorial.
01:47Until next time, see you guys.
Comments