3.Basic Settings
1.
Select [AI On-site Learning Management] from the submenu.
2.
Check the checkbox for the camera for which you want to configure this product.
This product is set by one camera. Check only one check box on the camera.
3.
Select [The entire process is carried out in a camera-connected environment].
4.
Click [Display].
-
[AI On-site Learning Settings] is displayed on a separate screen.
1.
Follow the guidance to make basic settings for the product
-
Each item has a symbol. Click this symbol to display the procedure in the lower left corner of the screen.

3.1Select the Detection Function
Detector 1 is set on the Guidance screen.

First, select the [On-site Learning Function] you want to use and the object you want to be detected.
When you check [Check this to set two or more objects or functions.], the screen display switches and settings can be made for up to 5 detection objects. Please refer to the following for the setting procedure.
→4.2　Set Detection Object
1.
Select [On-site Learning Function].
・
[Add a detection object]
-
Adds a new detection object.
-
The object name can be set in [Set Name].
-
If you check [Improve false detection do at the same time] checkbox, [Improve false detection] will be performed at the same time in addition to [Add a object]. You can select [Human], [Vehicle], or [Bicycle] as the object to be improved.
-
If [Improve false detection do at the same time] is checked,[Improve false detection] will be performed at the same time in addition to [Add a detection object]. You can select [Human], [Vehicle] or [Bicycle] as the object to be improved.
・
[Improve false detection]
-
Prevents false detection of humans, vehicles, or bicycles.
-
The target for improvements can be selected from [Human], [Vehicle] and [Bicycle].
・
[Improve missed detection]
-
Reduces missed detection of humans, vehicles, or bicycles.
-
The target for improvements can be selected from [Human], [Vehicle] and [Bicycle].
2.
Set the name of the object.
　
If [Add a object] is selected for [On-site Learning Function], set the name of the detection object. The name can be up to 20 single-byte characters.
3.
Select the detection object you want to improve.
　
If [Improve false detection] or [Improve missed detection] is selected for [On-site Learning Function], or if you have selected [Add a object] and checked [Improve false detection do at the same time], select the detection object you want to improve.

3.2Saving the Learning Image
Save the image to be used for learning.

You can save images directly from a pre-installed camera or from a JPEG/PNG file on your PC as learning images.

To take pictures directly from a pre-installed camera, select [Manually save camera image] or [Auto save camera image].
[Manually save camera image] allows you to manually save still images each time.
In [Auto save camera image], when you specify [Save interval] and [Number of save images], the camera automatically saves one still image for [Number of save images] at the specified [Save interval].

Select [Upload images from PC] to save jpg/png files on your computer as learning images.

3.2.1Saving camera images manually

1.
Select [Saving learning images].
-
Select [Manually save camera image].
2.
Click the save button.
-
One still image is shot.

3.2.2Saving camera images automatically
1.
Select [Saving learning images].
-
Select [Auto save camera image].
2.
Set the [Image saving interval].
-
Select from [10 s], [20 s], [30 s], [40 s], [50 s], [1min], [5min], [10min], [15min], [30min], [60min], and [When nothing detected].
-
When [When nothing detected] is selected, only images that could not be detected are efficiently saved.
3.
Select [Number of save images]
-
Select from [10], [20],..., [90], [100], [150], and [200].
4.
Click [Start]
-
To stop auto saving, click [Stop].
-
During auto save, the progress is displayed beside the [Stop].

3.2.3Upload images from PC
1.
Select [Saving the learning image].
-
Select [Upload images from PC].
2.
Click [Browse].
-
The Add File screen is displayed.
3.
Select the files you want to save as learning images, and then click [Open].
-
Saving starts.
-
The progress is displayed while the file is being saved.

・The learning image formats are compatible with.jpg and.png.


・The resolution of the image that can be saved as a learning image is between 640pixels and 3840pixels.


・It is recommended that images saved as learning images be acquired at the same angle of view as during operation.


・Up to 1000 learning images can be saved. However, depending on the resolution of the learning image, you may not be able to save up to 1,000 images.


・The percentage usage of the storage area for learned images can be checked on [Maintenance] window.


→4.8　Maintenance

・Learning images can also be learned in non-camera environment. Please refer to the following for details on learning in non-camera environment.


→7.2　Stop the work in a camera-connected environment and continue in a non-camera environment

3.3Set bounding boxes and learn
You can draw the bounding box you want the AI to learn in the learning image.
1.
Click [Display]
-
[Perform Learning Screen] is displayed.
2.
Click the thumbnail of image you want to use to set the bounding box.
-
[Learning Frame Settings Screen] is displayed.
When [Show the detection frame] is selected, the detection results for human/vehicle/bicycle/learned objects are displayed in gray. The detection result of human/vehicle/bicycle/learned objects exceeds the each detection thresholds. See below for detection thresholds.
→5.　Demo Screen
3.
Select [Add Bounding box] in [Mode] to draw a bounding box on the image.
-
You can set the bounding box as a rectangle by dragging it on the image.
-
Right-clicking inside a learning frame brings up a menu in the "Learning Frame Settings" window. Select [Delete] in the menu to delete the right-clicked bounding box. Select [Change to (the name of detected object)] in the menu to change the bounding box to that of the selected detected object. The same operation can be performed when [Edit Bounding box] is selected in [Mode].
4.
Select [Edit Bounding box] in [Mode] to edit a bounding box on the image.
-
Clicking on a drawn bounding box allows you to change the size and position, or delete it by clicking on the X icon.
-
Clicking the Icons for the detection results (gray frames) of human/vehicle/bicycle/learned objects will set the selected detection object to the bounding box.
-
The gray frames of human/vehicle/bicycle/learned objects can be shown or hidden with the "Show/hide the gray frame" checkbox.

・If the size of the learning frame you draw is too small or too large, the size is automatically adjusted. See the following for the minimum and maximum sizes.


→1.3　Camera Installation Conditions

・The detection result of human/vehicle/bicycle/learned object is displayed by checking [Show the detection frame] in the [Perform Learning Screen].


・You can set up to 100 bounding box per image.
5.
To set the learning area for the learning image, press [Setting up a learning area].
-
[Setting up a learning area] screen will be displayed in a separate window.
-
If the detected object is small, it can be easily detected by setting a learning area and limiting the detection range.
When the learning area is not set, it is difficult to detect a forklift in the back of the screen.
When a learning area is set up, it is easier to detect a forklift in the back of the screen.
The learning area is set up as follows.
・
Set the learning area on the image. By selecting the start and end points, the learning area is set in a rectangle.
Initially, the entire area is set as learning area and one area can be specified.
・
While selecting the start and end points of the learning area, the aspect ratio of the learning area is fixed to the same aspect ratio as the learning image.
In addition, the maximum size of the learning area that can be set is indicated by a dotted line.
・
When set, unlearned areas will be darkened.
・
To delete, press [Initialization].

・The learning area should be set so that the bounding boxes are within the learning area. Please set the learning area so that the bounding boxes to be learned fit within the learning area.



・Check the [Display the Size Checker] to check the minimum and maximum size of the bounding boxes that can be learned within the learning area you have set. When checked, the minimum size is indicated by a red frame and the maximum size by a green frame.


　Set the learning area so that the object you want to detect is larger than the red frame and smaller than the green frame.
6.
After completing the learning area settings, press [Set].
-
[Setting up a learning area] screen will close and return to the [bounging box Settings Screen].
-
In the [bounding box Settings Screen], an information mark will appear in the upper left corner of the bounding box if the bounding box is not the appropriate size for the learning area you have set. If you would like to make the bounding box for learning, select [Edit Bounding box] in [Mode] and adjust the size of the bounding box.
7.
Press [Next] to set the bounding box for the learning images.
-
The setting screen for the next image is displayed.
8.
When you have set all the learning images, click [Set].
-
[Learning Frame Setting Screen] closes, and you sre returned to the [Perform Learning Screen].
The [Set bounding boxes and learn] process can be interrupted halfway and continued in the camera unconnected environment. See the chapter below.
→7.2　Stop the work in a camera-connected environment and continue in a non-camera environment
9.
Select the image you want to use for learning
-
Select the check box at the bottom left of the thumbnails of the target images as shown.
-
Learning requires at least 10 images.You can learn up to 200 images.
-
The maximum number of bounding boxes that can be included in learning images are 1000.

・When bounding boxes are set, the check boxes at the bottom left of the thumbnails are automatically checked.


・When a learning area is set up, images with an aspect ratio different from the camera's image capture mode cannot be used for learning.
10.
Click the [Start Learning] to execute learning.
-
When learning is completed, the confirmation screen is displayed.

3.4Checking the Learning Results
・
Check the learning accuracy.
-
To check the image with live video, click [Display] on [Demo screen] to open the demonstration screen. See the following for information on the demonstration screen.
→5.　Demo Screen
-
To check images already shot, click [Display] on [Simulation screen] to open the simulation screen. See the following for more information on the simulation screen.
→6.　Simulation screen