The requirements for cameras that will work with the Neurocounter module are shown in the following table.
Video stream from camera | | Camera | - Resolution: at least 640x360 pixelsCorrect operation . It is also not guaranteed if a fish-eye lens is used
- Fps: not less than 2 (4 or more is recommended)
|
| Lightning | - The objects should be visually separated from the background and from each other
|
Scene and camera angle | - The camera should be installed at an angle at a height of at least 3 meters, or on the ceiling, or facing strictly down (for counting visitors)
|
Object images | The minimum width or height of the object should be at least 5% of the width/height of the frame- recommended to use a resolution higher than 1920×1080, since higher resolution does not increase the detection quality, but significantly increases the consumption of resources. The optimal resolution for solving typical tasks is 1280x720 (see Example of configuring Neurocounter for solving typical task).
- The frame rate per second in the video stream from the camera is at least 8 for solving typical task.
- Both colorless (gray) and color images
|
| Lighting | - Lighting in the scene is at least 50 lux per square meter. In conditions of insufficient or excessive lighting (night or light-striking), stable operation of the video analytics is not guaranteed.
- There are no abrupt changes in lighting
|
Scene and camera angle | - Moving objects are visually separable from each other.
- The background is mostly static and does not change abruptly.
- Moving objects are minimally obscured by static objects in the scene (columns, trees, etc.).
- The analyzed scene does not have reflective surfaces and sharp shadows from moving objects. If present, they should be masked.
- Camera shake does not result in image offsets greater than 1% of the frame size
|
Object images | Objects image requirements for Neurocounter |
| Note |
|---|
|
Correct operation of the neural counter is not guaranteed when using a fisheye lens. |
