# Depth camera object recognition

**URL:** <https://forum.opencv.org/t/depth-camera-object-recognition/15844>\
**Category:** Python\
**Created:** [December 14, 2023, 12:06pm UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844 "2023-12-14T12:06:21Z")\
**Posts on this page:** 8\
**Page:** 1

<div class="post-metadata">

**Author:** ![MobileRobotics](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/mobilerobotics/32/8060_2.png) [@MobileRobotics](https://forum.opencv.org/u/MobileRobotics)\
**Post date:** [December 14, 2023, 12:06pm UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/1 "2023-12-14T12:06:21Z")

</div>

I’m trying to recognize fruit types using a depth camera. For example, distinguish between bananas, apples and oranges. How should depth information be used without AI? Oranges and apples have similar colors, so I would like to use depth information.

---

<div class="post-metadata">

**Author:** ![crackwitz](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/crackwitz/32/14_2.png) [@crackwitz](https://forum.opencv.org/u/crackwitz)\
**Post date:** [December 14, 2023, 12:08pm UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/2 "2023-12-14T12:08:35Z")

</div>

> [@MobileRobotics](#):
>
> without AI

why not?

what level of machine learning are you allowed to use for this assignment? why is this limited?

---

<div class="post-metadata">

**Author:** ![MobileRobotics](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/mobilerobotics/32/8060_2.png) [@MobileRobotics](https://forum.opencv.org/u/MobileRobotics)\
**Post date:** [December 14, 2023, 12:53pm UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/3 "2023-12-14T12:53:26Z")

</div>

Machine learning can be used, but it is difficult because it requires learning cases in a short period of time.

---

<div class="post-metadata">

**Author:** ![berak](https://avatars.discourse-cdn.com/v4/letter/b/85f322/32.png) [@berak](https://forum.opencv.org/u/berak)\
**Post date:** [December 14, 2023, 1:25pm UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/4 "2023-12-14T13:25:58Z")

</div>

> [@MobileRobotics](#):
>
> Oranges and apples have similar colors, so I would like to use depth information.

and what would you get from the depth information ? size ? distance ?  
i cannot see anything useful for recognition here

---

<div class="post-metadata">

**Author:** ![MobileRobotics](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/mobilerobotics/32/8060_2.png) [@MobileRobotics](https://forum.opencv.org/u/MobileRobotics)\
**Post date:** [December 14, 2023, 11:04pm UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/5 "2023-12-14T23:04:12Z")

</div>

Isn’t depth information usually used when recognizing objects?

---

<div class="post-metadata">

**Author:** ![crackwitz](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/crackwitz/32/14_2.png) [@crackwitz](https://forum.opencv.org/u/crackwitz)\
**Post date:** [December 15, 2023, 8:47am UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/6 "2023-12-15T08:47:14Z")

</div>

it can be, but usually not. the usual models just work with color pictures. color pictures are easy to get. depth information is rarely available.

---

<div class="post-metadata">

**Author:** ![MobileRobotics](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/mobilerobotics/32/8060_2.png) [@MobileRobotics](https://forum.opencv.org/u/MobileRobotics)\
**Post date:** [December 19, 2023, 12:41am UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/8 "2023-12-19T00:41:55Z")

</div>

I can get the distance value for each pixel from the camera. Exactly what algorithm should be used to obtain object areas or features?

---

<div class="post-metadata">

**Author:** ![jsxyhelu](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/jsxyhelu/32/194_2.png) [@jsxyhelu](https://forum.opencv.org/u/jsxyhelu)\
**Post date:** [December 20, 2023, 7:23am UTC](https://forum.opencv.org/t/depth-camera-object-recognition/15844/9 "2023-12-20T07:23:59Z")

</div>

may be not,some picutres may be helpful
