What causes stretching distortion on a 3D camera?
Source:Shenzhen Kai Mo Rui Electronic Technology Co. LTD2026-07-31
On logistics lines, the most common issue with line laser 3D cameras isn't necessarily "no image"—rather, it's:
The image is stretched or compressed;
Abnormal contour spacing;
The point cloud is misaligned;
The collection rhythm doesn't match the conveyor line speed.
When encountering this type of issue, many people first suspect the algorithm, calibration, exposure, or device accuracy.

But in online laser applications, the first thing that should be checked is often:
Encoder, photoelectric triggering, and Line configuration.
Why does an encoder malfunction cause image stretching?
Ordinary cameras typically “trigger once and take one picture.”
A line laser 3D camera needs to move along with the conveyor belt, continuously acquiring profiles one by one and then stitching them together to form a complete 3D image.
Among them:
EncoderTell the device how much the conveyor line has moved;
Optoelectronic sensorTell the device when the object arrives.
If the encoder pulses are too few, too many, in the wrong direction, or fail to enter the device, it will result in abnormal spacing between contour segments, ultimately manifesting as image stretching, compression, or misalignment.
First, confirm the camera's I/O type.
The number of I/O ports may vary depending on the model of line-scan camera.
Some devices have only 1 pair of input and output, while others support 3 pairs of input and output.
For devices with 3 pairs of input and output, the recommendation is:
Line 3 and Line 6 connect to the encoder; Line 0 connects to the photoelectric trigger.
Please note that the Line configuration in the software must be exactly identical to the on-site wiring.
Although the encoder is actually connected to Line 3 and Line 6, the software has assigned Line 0 as the encoder input, so naturally the device cannot synchronize properly.
Kubler Encoder Wiring Precautions
If using a Kubler encoder, model:
8.KIS40.1342.1024
Refer to the 12 V common power supply connection method below.
The encoder outputs a pulse signal.
For every segment the conveyor belt moves, the encoder outputs a corresponding pulse, and the line laser adjusts its contour acquisition timing based on these pulses.
On-site key inspections:
Does the encoder have pulses entering the device?
Are phases A and B oriented correctly?
Is the pulse count continuous and stable?
Is the encoder resolution and pulse equivalent configured correctly?
Does the software input for Line match the wiring?
If pulses are lost or parameter settings are incorrect, the point cloud may become stretched or compressed.
How do you connect PNP photoelectric sensors?
The function of a photoelectric sensor is to inform the device that “an object has arrived.”
If using a PNP photoelectric sensor, refer to the wiring diagram below.
When the PNP phototransistor is activated, it outputs a high level.
There are three points that need to be confirmed on-site:
Which line does the fiber-optic cable connect to?
Is the signal ground or common terminal connected correctly?
Are the photoelectric power supply and the line laser power supply processed according to the diagram?
A change in the photoelectric indicator does not necessarily mean that the signal has entered the camera.
It’s best to directly observe in the software whether the input status of Line 0 changes synchronously.
Why should NPN optoelectronics be given special attention?
NPN inputs are more likely to be connected incorrectly in multi-channel signal scenarios.
One type of connection method is described in the material:
COMConnect the fiber optic cable;
SignalINConnect to the positive terminal of 12 V.
However, this method only supports a single input channel and cannot connect to an encoder simultaneously.
If both an encoder and an NPN photoelectric sensor are to be used on-site, it is recommended to adopt a multi-input method with pull-up resistors.
The pull-up resistor ensures that the input remains at a stable high level when no action is occurring, and then pulls the signal low when the NPN transistor is activated.
This can reduce:
Input state drift;
Random false triggering;
The encoder count is unstable;
Multiple signals interfere with each other.
Image stretching—recommended to troubleshoot in this order.
Don't jump straight into tuning the algorithm; it's recommended to check in the following order:
Confirm whether the encoder is sending pulses into the device.
Confirm that the encoder direction is correct;
Confirm whether the photoelectric trigger has entered Line 0.
Confirm that the software Line configuration matches the on-site wiring.
Finally, recheck the exposure, calibration, point cloud, and algorithm.
If the encoder and the photoelectric sensor aren't synchronized, no matter how much you fine-tune the algorithm later, you’ll just keep making repeated adjustments based on incorrect data.
Also, please note: Just because there’s a “trigger” doesn’t mean everything is normal.
The fact that photoelectric sensors can be triggered only indicates that the object-position signal has already been input into the device.
Only when the encoder pulses are correct can we be sure that the acquisition rhythm is truly synchronized with the conveyor's motion.
If low-speed operation is normal but high-speed stretching occurs, continue to perform further checks:
Encoder pulse frequency;
Camera contour acquisition frequency;
Exposure time;
Is there pulse loss?
Is the conveyor belt slipping?
Is the pulse equivalent setting correct?
Summary
When the image from a line laser 3D camera shows stretching, compression, or misalignment, don't immediately suspect the algorithm.
Should be checked first:
Does the encoder have pulses? Is the photoelectric sensor triggered correctly? Does the software’s Line configuration match the wiring?
One-sentence summary:
The key to line lasers isn't whether you can take the picture—it's whether the encoder, photoelectric sensors, and conveyor belt are truly synchronized. When stretching the image, first check the wiring; only then talk about the algorithm.
Related News
Summary of Algorithms for Defect Inspection in Machine Vision
2026-07-31Depth of Field: Concept, Calculation and Influencing Factors
2026-07-31What causes stretching distortion on a 3D camera?
2026-07-31Starlight Function of Surveillance Cameras
2026-07-30Differences Between Starlight Cameras, Full-Color Cameras and Conventional Infrared Cameras
2026-07-30Industrial inspection camera—High-speed camera solution for modern manufacturing quality control
2026-07-30






+8613798538021