Log Description
1. Log Storage Locations
（1）Log storage locations for versions before PickWiz V1.6.0
~/kuawei_data/log/middleware/middleware.log Some backend system logs other than the detection workflow ~/kuawei_data/log/backend/backend.log Backend detection workflow logs ~/kuawei_data/log/PickLightMicroServer/PickLight_deepmodel.log Model Inference service logs ~/kuawei_data/log/glia/glia.log glia logs ~/kuawei_data/log/rlia/rlia.log rlia logs ~/.dexforce/log/pickwiz/app.*.log PickWiz frontend logs
（2）Log storage locations for versions after PickWiz V1.6.0
~/.dexforce/log/picklight/backend_sys/backend_sys.log Formerly middleware.log ~/.dexforce/log/picklight/backend_pipe/backend_pipe.log Formerly backend.log ~/.dexforce/log/picklight/backend_deepmodel/backend_deepmodel.log Formerly PickLight_deepmodel.log ~/.dexforce/log/glia/glia.log glia logs ~/.dexforce/log/rlia/rlia.lAog rlia logs ~/.dexforce/log/pickwiz/app.*.log PickWiz frontend logs
（3）Log storage locations for versions after PickWiz 1.7.1
~/.dexforce/log/pickwiz/PickWiz_INFO_年月日-时分秒.进程ID PickWiz frontend logs
2. Log Function Introduction
| Field | Description |
|---|---|
| Time | The time when the log was generated. PickWiz supports real-time log viewing. |
| Level | The level of the log content. DEBUG indicates all log information, including detailed debugging information used for troubleshooting; INFO indicates logs for normal system operation, including operation processes, alarm information, and error information; WARNING indicates alarm information, including alarm information and error information; ERROR indicates error information. You can filter log content by Level at the bottom. |
| Content | The specific content of the log |
| Clear | Clears all logs currently displayed on the page. The log storage files are not cleared. If you need to view logs that have been cleared from the page, refer to 1. Log Storage Locations to open the corresponding log storage file. |
| Copy | Use Ctrl+C/Ctrl+V on the keyboard or right-click and copy with the mouse to copy log information, as shown below. |
| Search | Enter information in the log search box and click the Search🔍 icon to search for the corresponding information in the logs, and you can quickly display entries one by one up and down. |
3. Log Content
3.1 Data Source
The beginning of the log specifies the data source input to the Workflow.
If Camera data is selected in Vision Calculation Configuration - Data Source, the Camera needs to be connected. PickWiz will send an image capture command to the Camera, and the Camera will acquire 2D images and depth images in real time and generate a Point Cloud image.
As shown below, after the Camera is successfully connected for the first time, network speed measurement is performed first, followed by actual image capture. The actual image capture takes 1.94 s. After the capture is completed, the Camera data acquired in real time is parsed, including the following parameters.
RGB image resolution: the size of the 2D image, calculated as h (height) * w (width), in pixels;
Depth image resolution: the resolution of the depth image, calculated as h (height) * w (width), in pixels;
Scene Point Cloud quantity: a Scene Point Cloud generated based on the 2D image, depth image, and Camera Intrinsic Parameter, then input into the Workflow;
Camera parameters: that is, the Camera Intrinsic Parameter;
Robot pose at image capture trigger: the pose of the Robot Tool;
Robot joint angles: a representation of the Robot pose, generally expressed using Euler Angles and selected in Robot configuration;
Image capture timestamp: the time when image capture starts;
Detection action trigger: if the detection workflow is triggered by PickWiz, it is the client; if the detection workflow is triggered by the Robot, it is the Robot;
Data source: Camera data or historical data. For instructions about data sources, refer to Vision Calculation Configuration Operation Guide
If historical data is selected in Vision Calculation Configuration - Data Source, the Camera does not need to be connected. The Camera data saved in the historical data is parsed directly, and this process takes 0 time, as shown below.
3.2 Start Detection
After the data is parsed, the detection pipeline starts.
3.3 Foreground Processing
Foreground processing preprocesses the Camera data input to the Workflow. If functions are added in Vision Parameters - 2D Preprocessing, the 2D preprocessing functions process the 2D image, depth image, and Scene Point Cloud at this node.
3.4 Bin Detection
If there are Scene Objects such as bins in the scenario, this node detects the bins.
3.5 Instance Segmentation
Instance Segmentation refers to segmenting instances from the scenario, corresponding to Vision Parameters - 2D Recognition - Instance Segmentation.
As shown below, 9 instances are segmented from the scenario, and Point Clouds for the 9 instances are generated according to the Mask.
3.6 Instance Filtering
Instance Filtering refers to filtering the segmented instances, corresponding to Vision Parameters - 2D Recognition - Instance Filtering.
As shown below, instances with Confidence below 0.8 are filtered out. The Confidence of all 9 instances is greater than 0.8, so no instances are filtered out.
3.7 Instance Sorting
Instance Sorting refers to grouping and sorting the filtered instances, and the grouping and sorting method corresponds to Vision Parameters - 2D Recognition - Instance Sorting.
3.8 Pose Estimation
This node specifies the method for calculating the workpiece pose, corresponding to Vision Parameters - 3D Calculation - Pose Estimation.
As shown below, 3D registration performs coarse matching and fine matching between the Scene Point Cloud and the Point Cloud template.
3.9 Pick Point Generation
This node specifies the Pick Point generation method. It takes the pose estimation results as input and outputs the Pick Points generated on each instance.
3.10 Pick Point Filtering
This node specifies the method for filtering Pick Points. It takes all generated Pick Points as input and outputs the filtered Pick Points, corresponding to Vision Parameters - Pick Point Processing - Pick Point Filtering.
3.11 Pick Point Sorting
This node specifies the Pick Point sorting strategy. It takes the Pick Points generated after pose estimation as input and outputs the grouped and sorted Pick Points. Picking is performed in this order, corresponding to Vision Parameters - Pick Point Processing - Pick Point Sorting.
3.12 Detection End
After detection ends, this node specifies the number of segmented instances, the number of instances returned after instance filtering, the number of instances matched by pose estimation, the number of instances returned after Pick Point filtering, and the time consumption.
3.13 Send to the Robot
PickWiz sends Pick Points to the Robot. If Vision Calculation Configuration - Number of Instances to Send is set to a value greater than 1, the Pick Points of multiple instances are sent to the Robot each time; if Vision Calculation Configuration - Number of Pick Points to Send is set to a value greater than 1, multiple Pick Points of one instance are sent to the Robot each time.
