An I 2 2.5D imaging system combining a conventional industrial camera with an intelligent docking station over a C bus

Abstract

The application discloses a kind of based on I 2 2.5D imaging system and method of conventional industrial camera combined with intelligent docking station of C bus.The system includes conventional industrial camera, intelligent docking station, multi-direction controllable light source assembly and host computer.Camera has I 2 C host interface, is connected with intelligent docking station by I 2 C bus;Docking station built-in MCU, I 2 C slave interface and multi-channel light source drive channel, support independent adjustment the brightness, switch and trigger mode of each light source;Host computer issues control command by I 2 C interface of camera SDK, and runs photometric stereo algorithm.The application utilizes existing I 2 C channel of camera as light source control bus, saves independent light source controller, with wiring simple, time sequence synchronization precision high, expansibility strong, soft and hard decoupling, upgrading cost low etc.Advantages, especially suitable for conventional 2D vision production line to 2.5D ability low-cost upgrade.

Description

Technical Field

[0001] This invention belongs to the field of industrial vision inspection technology, specifically relating to a method based on I 2 A 2.5D imaging system and method combining a conventional industrial camera and a smart docking station based on the C-bus is particularly suitable for scenarios where existing conventional 2D vision production lines can be upgraded at low cost to obtain surface topography information. Background Technology

[0002] In industrial vision 2.5D / 3D surface imaging applications, precise timing control and brightness adjustment of multi-directional light sources are crucial to image quality. Currently, the main collaborative solutions for cameras and light sources in conventional industrial vision systems fall into the following categories: (1) Independent Light Source Controller Solution: A dedicated light source controller (such as CCS, MORITEX, etc.) is used. Switch and brightness commands are sent from the PC via serial port / network port / digital I / O, and the controller then drives each light source. In this solution, the camera and the light source controller are independent of each other, and triggering synchronization requires additional wiring and coordination with the PLC / industrial computer.

[0003] (2) Camera GPIO direct drive solution: The output IO of the industrial camera is used in conjunction with an external power driver board to switch the light source on / off. This solution is limited by the number of camera GPIOs (generally only 2 to 4 channels) and current capability, making it difficult to adjust the brightness and control multiple light sources at the same time.

[0004] (3) Integrated light source camera solution: The light source and control circuit are directly integrated into the camera body. This solution has a high degree of camera customization. Once the light source type, wavelength and shape are fixed, it is difficult to switch them according to the object being measured on site.

[0005] (4) PLC centralized control scheme: The PLC uniformly schedules camera triggering and light source switching, and all light sources are directly connected to the PLC output module. This scheme has complex wiring, and the PLC's IO response delay is generally in the millisecond range, which is difficult to meet the high-speed light source switching requirements of the microsecond range.

[0006] (5) Most of the existing 2.5D photometric stereoscopic solutions use the above-mentioned independent controller or PLC solution, which limits the timing coordination, wiring complexity and expandability.

[0007] Existing solutions have drawbacks such as long control links, difficulty in timing coordination, complex wiring, poor scalability, difficulty in independently and finely adjusting brightness, poor decoupling from the camera, and high upgrade costs. Summary of the Invention

[0008] This invention provides a method based on I 2The 2.5D imaging system and method combining a conventional industrial camera with a smart expansion dock based on the C-bus aims to solve the problems of long light source control links, low timing synchronization accuracy, complex wiring, poor scalability, and high upgrade costs in existing 2.5D imaging systems.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a method based on I 2 A 2.5D imaging system combining a standard industrial camera with a smart docking station based on the C-bus, including: Conventional industrial camera, the conventional industrial camera having I 2 The C host interface can be used to call I via the software development kit (SDK). 2 C read / write transactions; The intelligent expansion dock, the intelligent expansion dock via I 2 The C-bus connects to the conventional industrial camera, and the intelligent expansion dock has a built-in main control unit (MCU) and I-band. 2 C slave interface and multiple light source drive channels; A multi-directional controllable light source assembly, which consists of N independently controllable light sources, each of which is electrically connected to the corresponding light source drive channel of the intelligent expansion dock. The host computer, which is communicatively connected to the conventional industrial camera, is used to call the camera SDK's I / O methods. 2 Light source control commands under the C interface.

[0010] Furthermore, the main control unit of the intelligent expansion dock is a microcontroller, and the multi-channel light source drive channel includes a PWM dimming circuit or a constant current DAC dimming circuit, which is used to independently adjust the brightness, on / off state and trigger mode of each light source.

[0011] Furthermore, a trigger synchronization line is connected between the conventional industrial camera and the intelligent expansion dock. The trigger synchronization line is used to directly transmit the camera's exposure trigger signal to the intelligent expansion dock, so as to achieve microsecond-level synchronization between the light source switch and the camera exposure.

[0012] Furthermore, the I of the intelligent expansion dock 2 The C slave interface is assigned a unique slave address and has a defined register mapping table, which includes a channel enable register, a brightness register, a trigger mode register, and a status read-only register.

[0013] Furthermore, the intelligent expansion dock is also equipped with a sequence template register for preloading the timing of multiple light sources, and the host computer starts the timing acquisition of the entire group of multi-directional light sources through a single trigger command.

[0014] Furthermore, the intelligent expansion dock also includes a power management module, which provides power to each light source at different voltage levels and has overcurrent protection and overtemperature protection circuits.

[0015] Furthermore, the host computer includes a photometric stereo algorithm module, which is used to solve the surface normal vector, albedo map and relative height map based on the multi-directional illumination image sequence, and output a 2.5D multi-channel image.

[0016] Furthermore, the number N of the multi-directional controllable light source assembly is 4, 6, 8, 12 or 16, and each light source is symmetrically distributed around the camera optical axis.

[0017] A type of I-based 2 A 2.5D imaging method combining a conventional industrial camera with a smart docking station on a C-bus includes the following steps: S1, The host computer uses the camera SDK's I... 2 The C interface reads the identification information of the smart docking station and completes the system handshake and initialization; S2. Acquire images of the calibration object under multi-directional light sources using a camera, and calibrate the direction vector and relative light intensity of each light source. S3. The host computer writes the lighting order and duration of multiple light sources to the sequence template register of the smart expansion dock and preloads the timing template. S4. The host computer sends a single trigger command to the smart expansion dock. The smart expansion dock turns on each light source in sequence according to the timing template and triggers the camera exposure simultaneously to collect a set of multi-directional lighting images. S5. The host computer uses a photometric stereo algorithm to solve the surface normal vector and albedo of each pixel based on the acquired image sequence, and obtains the relative height map through integration. S6 outputs grayscale image, normal vector image, albedo image and relative height image to complete 2.5D imaging.

[0018] Furthermore, the method for solving the normal vector in step S5 includes: constructing the observation matrix based on the Lambert diffuse reflection model, solving the intermediate vector by the least squares method, and then separating the albedo and the unit normal vector; the integration method adopts the Frankot-Chellappa frequency domain integration method or the Poisson solver.

[0019] Compared with the prior art, the beneficial effects of the present invention are: (1) Extremely simple wiring: Only one I cable is needed between the camera and the expansion dock. 2 The C-cable (SDA / SCL / GND / Power, 4 wires) eliminates the need for separate USB / serial / network ports and corresponding power cables compared to traditional independent controller solutions.

[0020] (2) Precise timing: The camera Strobe signal directly enters the hardware timer of the docking station MCU, and the synchronization accuracy of exposure and light source switching is <2μs, which is far superior to PLC (millisecond level) and PC control schemes.

[0021] (3) Strong expandability: The docking station can support 4 / 6 / 8 / 12 / 16 light sources, each with independent PWM / DAC dimming (12bit), and the SDK can dynamically adjust the brightness online.

[0022] (4) Decoupling of hardware and software, flexible combination: The camera, expansion dock and light source are decoupled, and different brands of cameras, different number of light sources and different forms of light sources (ring, strip, coaxial, etc.) can be freely combined according to the needs of the site.

[0023] (5) Zero camera modification and low upgrade cost: The existing conventional 2D vision production line only needs to add an expansion dock and multi-directional light source, and the software layer can be upgraded to obtain 2.5D capability.

[0024] (6) Health monitoring: Light source current / temperature is monitored via I 2 C-level feedback can provide lifespan warnings and fault location, improving production line stability. Attached Figure Description

[0025] Figure 1 This is a block diagram of the overall system structure of the present invention.

[0026] Figure 2 This is a block diagram of the internal circuitry of the intelligent expansion dock of the present invention.

[0027] Figure 3 This invention I 2 C register mapping representation intent.

[0028] Figure 4 This is a flowchart of the 2.5D imaging method of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Example 1: This example provides a method based on I 2 A 2.5D imaging system combining a conventional industrial camera with a smart docking station using the C-bus.

[0031] System Composition Conventional industrial cameras: Select those supporting I 2A monochrome area scan camera with a GigE interface for C-type pass-through, whose camera SDK provides CamI... 2 The C_Write / Read interface. The camera exposes SDA, SCL, and Strobe signal lines.

[0032] Intelligent expansion dock: The internal main control MCU adopts STM32F103, I 2 The slave address (C) is set to 0x42. It is configured with 8 light source drive channels, each using MOSFET switching + PWM dimming (10kHz frequency, 12-bit resolution). The expansion dock also includes a trigger input (for connecting to the camera Strobe) and a power management module (12V / 24V output).

[0033] Multi-directional controllable light source: 8 LED ring light sources, evenly distributed around the camera optical axis (one source every 45°), with a pitch angle of 45°.

[0034] Host computer: Industrial control computer, running 2.5D algorithm software, communicating with the camera via GigE.

[0035] Hardware connection The camera is connected to the expansion dock via a 4-core cable (SDA, SCL, GND, power), plus an additional Strobe trigger cable. Each light source is connected to the expansion dock's output port via a 2-core cable.

[0036] I 2 C protocol Define register mapping: 0x00-0x01 is ID / VERSION read-only; 0x10+i is the enable of the i-th channel (1 byte); 0x20+i is the brightness of the i-th channel (2 bytes, 0-4095); 0x30+i is the trigger mode (1 byte); 0x40 onwards is the timing template register (N×2 bytes, storing the lighting sequence and duration respectively); 0x50 is the trigger start register (writing 1 starts a sequence); 0x60+i is the status read-only register (2 bytes, current + temperature).

[0037] 2.5D Imaging Process: Step S1: System Initialization and Handshake The host computer calls the camera SDK via I 2 C reads the ID / VERSION register of the docking station to confirm that the docking station model and firmware version are compatible; it also reads capability descriptions such as the number of channels N and maximum brightness.

[0038] Step S2: Light source calibration (one-time) A calibration sphere (a specular / diffuse sphere of known radius) is placed at the center of the field of view. N light sources are turned on sequentially (only one is lit at a time), and the camera captures N calibration images. The host computer algorithm module calculates the light source direction vector l_i∈R based on the position of the specular highlight on the sphere. 3 The relative light intensity e_i is estimated based on the diffuse reflection area. The calibration results are stored on the host computer algorithm side (or can be written back to the docking station EEPROM).

[0039] Step S3: Sequence Configuration The host computer loads the 2.5D light source timing sequence via `Dock_LoadSequence`: in the first frame, only the first light source is lit; in the second frame, only the second light source is lit; and so on. The brightness of each light source is pre-adjusted according to the surface reflection characteristics of the object being measured (typically, channel compensation can be used when the difference is greater than 2 times).

[0040] Step S4: Synchronous Trigger Acquisition The host computer calls `Dock_TriggerOnce` to start the sequence, and the docking station MCU follows the timing sequence: Turn on the i-th light source ~5μs before the exposure of the i-th frame; The camera exposure is triggered by the Strobe signal; After the exposure is complete, turn off the i-th light source; Once the camera is ready, proceed to frame i+1.

[0041] Once the entire set of N frames (typically N=4, t≈40~80ms) has been acquired, the camera will transmit the N images back to the host computer.

[0042] Step S5: Solve using the photometric stereo algorithm Solve for N observations stacked per pixel:

[0043]

[0044] Step S6: Integrate the normal vector to the relative height map Depend on Derivation of gradient field The relative height map Z(x,y) is recovered using the Frankot-Chellappa frequency domain integral or the Poisson solver.

[0045] Step S7: 2.5D Multi-channel Output Output grayscale image + normal vector image + albedo image + relative height Figure 4 The channel is used for downstream defect detection, OCR, and other applications.

[0046] Step S8: State Closure The SDK periodically communicates via I2 C reads the `CH_i_STATUS` register of the expansion dock to monitor the current and temperature of each light source, and promptly detects and alarms light source faults.

[0047] Example 2: This example is used for upgrading an outdated 2D vision production line. The production line originally had a conventional 2D camera, which now needs to detect minute indentations on the surface of metal stamping parts. The user purchased the intelligent expansion dock and four directional light sources (0°, 90°, 180°, 270°) of this invention. Following the wiring and calibration steps of Example 1, without replacing the camera, only the expansion dock and light sources were added. By upgrading the software and calling the 2.5D algorithm, the indentation defects were successfully detected, and the detection rate was increased from less than 60% to over 95%.

[0048] Example 3: This example is used in a flexible production line that needs to inspect different products (phone frames, PCB boards, injection molded parts). The same camera and docking station can be used to quickly change different types of light source components (such as ring light sources, strip light sources, and coaxial light sources) depending on the object being tested. After changing the light source, the light source direction is recalibrated (step S2) for normal operation; the rest of the system requires no modification.

[0049] Example 4: In this example, the expansion dock supports 12 light sources for higher precision photometric stereo reconstruction. After acquiring 12 multi-directional illumination images, the accuracy of the normal vector calculation is further improved, and it still has a good response to extremely weak scratches (depth <5μm).

Claims

1. A method based on I 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... include: Conventional industrial camera, the conventional industrial camera having I 2 The C host interface can be called via a software development kit. 2 C read / write transactions; The intelligent expansion dock, the intelligent expansion dock via I 2 The C-bus connects to the conventional industrial camera, and the intelligent expansion dock has a built-in main control unit and I-band. 2 C slave interface and multiple light source drive channels; A multi-directional controllable light source assembly, which consists of N independently controllable light sources, each of which is electrically connected to the corresponding light source drive channel of the intelligent expansion dock. The host computer, which is communicatively connected to the conventional industrial camera, is used to call the camera's software development kit (SDK) I / O function. 2 Light source control commands under the C interface.

2. A method based on I according to claim 1 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... The main control unit of the intelligent expansion dock is a microcontroller, and the multi-channel light source drive channel includes a PWM dimming circuit or a constant current DAC dimming circuit, which is used to independently adjust the brightness, on / off state and trigger mode of each light source.

3. A method based on I according to claim 1 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... A trigger synchronization line is also connected between the conventional industrial camera and the intelligent expansion dock. The trigger synchronization line is used to directly transmit the camera's exposure trigger signal to the intelligent expansion dock to achieve microsecond-level synchronization between the light source switch and the camera exposure.

4. A method based on I according to claim 1 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... The intelligent expansion dock I 2 The C slave interface is assigned a unique slave address and has a defined register mapping table, which includes a channel enable register, a brightness register, a trigger mode register, and a status read-only register.

5. A method based on I according to claim 4 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... The intelligent expansion dock is also equipped with a sequence template register for preloading the timing of multiple light sources. The host computer starts the timing acquisition of the entire group of multi-directional light sources through a single trigger command.

6. A method based on I according to claim 1 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... The intelligent expansion dock also includes a power management module, which provides power to each light source at different voltage levels and has overcurrent and overtemperature protection circuits.

7. A method based on I according to claim 1 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... The host computer includes a photometric stereo algorithm module, which is used to solve the surface normal vector, albedo map and relative height map based on the multi-directional illumination image sequence, and output a 2.5D multi-channel image.

8. A method based on I according to claim 1 2 A 2.5D imaging system combining a conventional industrial camera with a smart expansion dock and a C-bus, characterized in that... The number of light sources N in the multi-directional controllable light source assembly is 4, 6, 8, 12 or 16, and each light source is symmetrically distributed with the camera optical axis as the center.

9. A method based on any one of claims 1 to 8, using I 2 The 2.5D imaging method of a 2.5D imaging system combining a conventional industrial camera with a smart expansion dock on a C-bus is characterized by, Includes the following steps: S1, The host computer uses the camera software development kit (I) 2 The C interface reads the identification information of the smart docking station and completes the system handshake and initialization; S2. Acquire images of the calibration object under multi-directional light sources using a camera, and calibrate the direction vector and relative light intensity of each light source. S3. The host computer writes the lighting order and duration of multiple light sources to the sequence template register of the smart expansion dock and preloads the timing template. S4. The host computer sends a single trigger command to the smart expansion dock. The smart expansion dock turns on each light source in sequence according to the timing template and triggers the camera exposure simultaneously to collect a set of multi-directional lighting images. S5. The host computer uses a photometric stereo algorithm to solve the surface normal vector and albedo of each pixel based on the acquired image sequence, and obtains the relative height map through integration. S6 outputs grayscale image, normal vector image, albedo image and relative height image to complete 2.5D imaging.

10. A method based on I according to claim 9 2 A 2.5D imaging method combining a conventional industrial camera with a smart docking station using the C-bus is characterized by: The method for solving the normal vector in step S5 includes: constructing the observation matrix based on the Lambert diffuse reflection model, solving the intermediate vector by the least squares method, and then separating the albedo and the unit normal vector; the integration method uses the Frankot-Chellappa frequency domain integration method or the Poisson solver.

Images