2D Target Object Recognition Tool
This section introduces the main functions and typical application scenarios of the 2D Target Object Recognition Tool.
Introduction
The 2D Target Object Recognition tool is a visual debugging tool that integrates common visual recognition processing functions, supporting four major application scenarios: positioning and picking, placement correction, error-proofing check, and information reading.
-
Positioning and Picking: Recognize the pose of same-type target objects in the XOY plane using a 2D camera (Z-axis height should be consistent). Suitable for picking non-stacked objects from trays and objects lying flat on conveyor belts.
-
Placement Correction: Before placing a target object, capture an image with a 2D camera to recognize its current position, and perform correction based on the recognition result to achieve precise placement. Suitable for scenarios such as assembly where high placement accuracy is required.
-
Error-proofing Check: Verify the status of target objects using 2D camera images to confirm they meet expected conditions, preventing errors such as missing objects, reversed orientation, or wrong objects. Suitable for scenarios including orientation check, presence check, deformation check, misalignment and skew check, and quantity counting.
-
Information Reading: Read 1D or 2D barcodes, or character information on target objects using a 2D camera to acquire key identification data. Suitable for material traceability and model identification scenarios.
Reading Guidance
The configuration workflow of the 2D Target Object Recognition tool varies by scenario. Please read the following sections for detailed instructions.
Prerequisites:
-
2D camera calibration has been completed. If image distortion exists, distortion calibration must be performed first (applicable to all scenarios); positioning and picking and placement correction scenarios also require extrinsic parameter calibration.
-
The camera can capture images normally, and the image quality meets the requirements. To adjust camera parameters, please read 2D Camera Management to optimize the parameter settings of the currently connected camera.
Positioning and Picking
In the positioning and picking scenario, target object recognition can be performed through two methods: 2D template matching and 2D Blob analysis.
-
2D template matching: If you need to search and locate target object features matching a template in a 2D image, select the 2D template matching recognition method. For detailed configuration, please read 2D Template Matching.
-
2D Blob analysis: If you need to detect Blobs in an image and filter targets based on geometric features, select the 2D Blob analysis recognition method. For detailed configuration, please read 2D Blob Analysis.
Placement Correction
In the placement correction scenario, target object recognition can be performed for both fixed and non-fixed placement targets.
-
Fixed placement target: If the position and orientation of the placement target remain unchanged during operation, select the fixed placement target option for target object recognition. For detailed configuration, please read Fixed Placement Target.
-
Non-fixed placement target: If the actual position or orientation of the placement target varies between runs, select the non-fixed placement target option for target object recognition. For detailed configuration, please read Non-fixed Placement Target.
| Before starting the specific configuration workflow for non-fixed placement target scenarios, please first configure the placement target recognition project. The specific recognition method can be flexibly designed based on actual requirements. After recognition is completed, the recognized placement target pose must be correctly saved to a global variable. For global variable configuration, please refer to Global Variable. |
Error-Proofing Check
In the error-proofing check scenario, target object recognition can be performed through four methods: binary classification, deformation classification, misalignment and tilt classification, and 2D template matching and counting.
-
Binary classification: If you need to determine the front-back orientation of a target object or whether a target object is present, select binary classification. For detailed configuration, please read Binary Classification.
-
Deformation classification: If you need to detect whether a target object has deformed, select deformation classification. For detailed configuration, please read Deformation Classification.
-
Misalignment and tilt classification: If you need to detect whether the placement position or orientation of a target object is abnormal, select misalignment and tilt classification. For detailed configuration, please read Misalignment and Tilt Classification.
-
2D template matching and counting: If you need to recognize target objects based on 2D template matching and count the number of target objects, select 2D template matching and counting. For detailed configuration, please read 2D Template Matching and Counting.
Information Reading
In the information reading scenario, target recognition can be performed through two methods: 1D/2D barcode recognition and optical character recognition.
-
1D/2D barcode recognition: If you need to read 1D or 2D barcode information on the target, select 1D/2D barcode recognition. For detailed configuration, please read 1D/2D Barcode Recognition.
-
Optical character recognition: If you need to recognize characters printed, engraved, or inkjet-marked on the target surface, select optical character recognition. For detailed configuration, please read Optical Character Recognition.