Camera Label Image Acquisition

## Camera Label Image Acquisition

![](/media/202601/2026-01-13_161942_0903030.26123893740286075.png)

## Module Principle

The essence of line scan camera label acquisition is to sequentially capture each image feature (label) according to the order of appearance of image features in "continuous images".

### Algorithm Workflow

<p><span>In the graphic and text inspection system of line scan cameras, since the image is generated in a </span><strong><span>line-by-line acquisition, infinite extension</span></strong><span> manner, deviations will occur after the encoder runs for a long time, causing the camera to fail to capture a complete label within the imaging height. Therefore, a </span><strong><span>label acquisition algorithm</span></strong><span> is needed to convert the continuous line scan image stream into </span><strong><span>structured, croppable, and identifiable image label regions</span></strong><span>, providing a foundation for subsequent OCR, pattern comparison, quality inspection, and data traceability.</span></p><p><span>From the perspective of engineering algorithms, the following explanations are made:</span></p><ul data-spread="false"><li><p><span>Dual-frame image stitching (Line → Strip → Virtual Image)</span></p></li><li><p><span>NCC template matching (or gap positioning algorithm)</span></p></li><li><p><span>Image cropping (label region generation)</span></p></li></ul>

<div class="cm-line">Algorithm Flowchart</div><div class="cm-line">    A[Line scan camera acquires single line image] --&gt; B[Line buffer write]</div><div class="cm-line">    B --&gt; C[Vertical image stitching]</div><div class="cm-line">    C --&gt; D[NCC template matching]</div><div class="cm-line">    D --&gt; E[Matching score curve Score(y)]</div><div class="cm-line">    E --&gt; F{Score &gt; Threshold?}</div><div class="cm-line">    F --&gt;|Yes| G[Enter label region]</div><div class="cm-line">    F --&gt;|No| H[Continue sliding]</div><div class="cm-line">    G --&gt; I[Stability determination m/n]</div><div class="cm-line">    I --&gt; J[Determine Start / End]</div><div class="cm-line">    J --&gt; K[Image cropping]</div><div class="cm-line">    K --&gt; L[Output label image]</div>

## Usage Scenarios
Provide label data source for graphic / printed label inspection.

## Usage Method
1. In VG3 software, drag a Camera Label Image Acquisition module.
2. Connect the line scan camera.
3. Create a product, set the label plate length and scale.
4. Create a matching template. For graphics and text with smaller label sizes, you can choose NCC matching.
For larger sizes, choose gap positioning algorithm.

## Feature Template
1. When choosing NCC matching, when framing the matching template, do not frame too large, just slightly larger than the label size (a few pixels, enough to display all label information).
2. When choosing gap positioning, when framing the matching template, it should be slightly larger than the label size (need to be able to display the gap of the entire label).

## Parameter Configuration Description
**Module Parameter Settings:**
![](/media/202601/block_param_1768275749.png)
Physical Camera:
1. Select the line scan camera to be connected.
2. Set the gain and exposure of the camera.
Simulation Camera:
1. Add pictures to be tested.
2. Set image format (color or monochrome).

**Product Parameter Settings:**
![](/media/202601/m_size_1768276091.png)
Plate Length: The actual height of the label to be collected (unit: mm).
Scale: The height (mm) corresponding to 1 vertical pixel in the image in the real physical world.

**Algorithm Parameter Settings:**
![](/media/202601/diwei_1768272474.png)
1. Camera Pixel Detection Range: The pixel position of the label in the picture.
2. Label Gray: Refers to the contrast gray of the label and the label background. If the label is darker than the background, select Dark, otherwise select Light.
3. Camera Background: Refers to the contrast gray of the picture background collected by the camera and the label. If the picture background is darker than the label, select Dark parameter, otherwise select Light.
4. Camera Background Gray Difference: This needs to evaluate the gray difference between the picture background and the label.

## Output Description
1. Running Result: Returns whether a target label is successfully collected.
2. Collected Image: Returns the cropped label image.
3. Image Name: Returns the name information of the cropped image.

## Precautions
When using line scan camera label image acquisition, please note: When setting the matching template, do not set the template too large, otherwise there is a risk of losing images in extreme cases. When using the gap positioning algorithm for large images, you can set the imaging height of the line scan camera (Product setting: Plate Length / Scale = Imaging Height) to be about 1.15 times the label. When using NCC matching, just set the height of the label directly.

Camera Label Image Acquisition

---

Click here to jump to the corresponding Chinese manual page: [Camera Label Image Acquisition](/doc/36)