Wiki

Analog Devices Wiki

English
English
简体中文
简体中文
日本語
日本語
Руccкий
Руccкий
This version (05 Dec 2022 05:28) was approved by Harshvardhan Bhatia.The Previously approved version (30 Nov 2022 19:37) is available.Diff

Table of Contents

EVAL-ADSD3175D Development

GitHub repo for SDK source code: ToF

Doxygen documentation (via the GitHub ToF repo): https://analogdevicesinc.github.io/ToF/

The SDK API is used to control the setup the camera and SDK, control the camera and get frames.

There are several examples provided with the SDK: ToF/tree/master/examples

The SDK uses the name space aditof.

Architecture

Requirements:

  • The SDK is put forward as a reference design for customers. Note, the SDK is under MIT license.
  • The SDK is also used in the EVAL-ADTF3175D evaluation platform.
  • Evaluation on Windows and Linux.
  • Run without the host on Linux embedded systems.

Case 1: SDK for use in EVAL-ADTF3175D

Device Embedded Linux Channel Host (Windows or Linux)
ADSD3500 ucv-app linked with SDK w/Adsd3500Sensorclass and partial depth compute stubs USB 3.0 UVC
* Protobuf is used for commands. 		User app linked with SDK w/UVC enabled and partial depth compute

Case 2: SDK for use in Embedded Linux SoC

Device Embedded Linux
ADSD3500 User app linked with SDK w/Adsd3500Sensorclass and partial depth compute

SDK Version

The following APIs can be used to retrieve version and Git details on the SDK in use.

  • aditof::getApiVersion(): The API version of the SDK.
  • aditof::getBranchVersion(): The branch name from which the SDK was built.
  • aditof::getCommitVersion(): The commit hash for the SDK.

Return Codes

The SDK uses the Status enum for return codes. See https://analogdevicesinc.github.io/ToF/namespaceaditof.html#a44f136ae2a9a546362d75a5d73443da9.

Dissecting the first-frame C++ Example

Note, it is expected the following example runs on the NXP directly - and not the host, being Windows or Desktop Linux.

Source code: ToF?master/examples/first-frame/main.cpp

Let's start by looking at some of the basic classes that are needed.

Step 1: Get List of Available Cameras

The following code block reads a list of cameras discovered by the SDK.

The System class is used to obtain the camera list by passing which results in a vector of Camera class objects. 

    System system;

    std::vector<std::shared_ptr<Camera>> cameras;
    system.getCameraList(cameras);
    if (cameras.empty()) {
        LOG(WARNING) << "No cameras found";
        return 0;
    }

    auto camera = cameras.front();

Step 2: Apply SDK Configuration File

In this step we specify the SDK configuration JSON file. 

    status = camera->setControl("initialization_config", configFile);
    if(status != Status::OK){
        LOG(ERROR) << "Failed to set control!";
        return 0;
    }

An example of the JSON file. 

{
"skip_network_cameras": "off",
"DEPTH_INI": "../config/RawToDepthAdsd3500_lrqmp.ini;../config/RawToDepthAdsd3500_lrmp.ini;../config/RawToDepthAdsd3500_srqmp.ini;../config/RawToDepthAdsd3500_srmp.ini",
"FPS": "10",
"FSYNC_MODE": "1"
}

Step 3: Initialize the Camera

Initialize the selected camera. 

    status = camera->initialize();
    if (status != Status::OK) {
        LOG(ERROR) << "Could not initialize camera!";
        return 0;
    }

Step 4 (optional): Display Version Details

This optional step returns and displays version details. These details are populated based on the ADI ToF eval platform. The user may need to populate these in a different manner in their platform. In which case investigate the CameraDetails class further. 

    aditof::CameraDetails cameraDetails;
    camera->getDetails(cameraDetails);

    LOG(INFO) << "SD card image version: " << cameraDetails.sdCardImageVersion;
    LOG(INFO) << "Kernel version: " << cameraDetails.kernelVersion;
    LOG(INFO) << "U-Boot version: " << cameraDetails.uBootVersion;

Step 5: Get Available Frame Types

Returns details on the frame information available from the camera. 

    std::vector<std::string> frameTypes;
    camera->getAvailableFrameTypes(frameTypes);
    if (frameTypes.empty()) {
        std::cout << "no frame type avaialble!";
        return 0;
    }

To understand what can be returned the please reference the definition of availableFrameTypes in adsd3500_sensor.h

For the Crosby, the frames are defined in the code block below. From this we can see the attached camera provides the following modes: lrqm, srqmp, lrmp and srmp. As well as the types of frames available for each mode and the sizes of the frames. 

    const std::vector<aditof::DepthSensorFrameType> availableFrameTypes = {
        {
            "lrqmp",
            {{"raw", 2560, 512},
             {"ir", 512, 512},
             {"xyz", 512, 512},
             {"depth", 512, 512},
             { "embedded_header",
               1,
               128 }},
            2560,
            512,
        },
        {
            "srqmp",
            {{"raw", 2560, 512},
             {"ir", 512, 512},
             {"xyz", 512, 512},
             {"depth", 512, 512},
             { "embedded_header",
               1,
               128 }},
            2560,
            512,
        },
        {
            "lrmp",
            {{"raw", 1024, 4096},
             {"ir", 1024, 1024},
             {"xyz", 1024, 1024},
             {"depth", 1024, 1024},
             { "embedded_header",
               1,
               128 }},
            1024,
            4096,
        },
        { "srmp",
          {{"raw", 1024, 4096},
           {"ir", 1024, 1024},
           {"xyz", 1024, 1024},
           {"depth", 1024, 1024},
           { "embedded_header",
             1,
             128 }},
          1024,
          4096,
        }};

For the definition of DepthSensorFrameType and DepthSensorFrameContent , see sensor_definitions.h.

Step 6: Select the Mode/FrameType

    status = camera->setFrameType("lrqmp");
    if (status != Status::OK) {
        LOG(ERROR) << "Could not set camera frame type!";
        return 0;
    }

Step 7: Start the Camera

    status = camera->start();
    if (status != Status::OK) {
        LOG(ERROR) << "Could not start the camera!";
        return 0;
    }

Step 8: Read a Frame from the Device

    aditof::Frame frame;

    status = camera->requestFrame(&frame);
    if (status != Status::OK) {
        LOG(ERROR) << "Could not request frame!";
        return 0;
    } else {
        LOG(INFO) << "succesfully requested frame!";
    }

Step 9: Get the Data from the Frame

Stage 0

  • Note, this is added for code simplification. Pay attention to Stages 1 and 2.
    save_frame(frame, "ir");

Stage 1

  • Using the frame retrieved in step 8, request a specific type of data from the frame.
  • Where the specific frame type data was seen in step 5: raw, ir, depth, …
Status save_frame(aditof::Frame& frame, std::string frameType){

    ...
    uint16_t *data1;
    ....

    status = frame.getData(frameType, &data1);
    if (status != Status::OK) {
        LOG(ERROR) << "Could not get frame data " + frameType + "!";
        return status;
    }

    ...

Stage 2 

    ...
    FrameDataDetails fDetails;
    ...
    frame.getDataDetails(frameType, fDetails);
    g.write((char*)data1, fDetails.width * fDetails.height * sizeof(uint16_t));
    ...
}

first-frame.py

The following section goes through important chunks of the python example.

ToF/tree/master/bindings/python/examples/first_frame_network

Specify IP of all connected cameras

Add all of the camera IPs in the list below 

 IP_addr = ['10.42.0.1']

Trigger mode

 # FSYNC TOGGLE MODE
 # 0 - ADSD3500 Fsync toggled by host command (press enter for next frame)
 # 1 - ADSD3500 Fsync toggled at framerate specified in json file
 # 2 - ADSD3500 Fsync disabled, pin set to HiZ
 fsync_toggle = 1

Initialize Cameras

Gets cameras based on specified IP 

 # status = system.getCameraList(cameras)
 status = system.getCameraListAtIp(camera, IP_addr[i])
 print("system.getCameraListAtIp()", status)
 cameras.append(camera[0])

Initialize Cameras

Initializes camera based on settings in user-defined json file. 

 cameras[i].setControl("initialization_config", "tof-viewer_config.json")
 status = cameras[i].initialize()
 print("camera"+IP_addr[i]+".initialize()", status)
 cameras[i].setControl("loadModuleData", "call")

Set Frame Type and start camera

 status = cameras[i].setFrameType("lrqmp")
 print("IP_addr[i].setFrameType()", status)
 status = cameras[i].start()
 print("camera"+ IP_addr[i]+".start()", status)

Toggle mode write

 cameras[i].adsd3500_set_toggle_mode(fsync_toggle)
 print("camera"+ IP_addr[i]+".adsd3500_set_toggle_mode()", status)

Trigger FSYNC (toggle_mode = 0)

If toggle mode is zero, this requires the user to trigger fsync for each frame 

 if fsync_toggle == 0:
    input('Press enter for next frame')
    status = cameras[i].adsd3500_toggle_fsync()
    print("camera"+ IP_addr[i]+".adsd3500_toggle_fsync()", status)

Capture frame

Call request frame by passing a tof.Frame object. 

 frames.append(tof.Frame())
 status = camera.requestFrame(frames[i])
 print("camera"+str(i)+".requestFrame()", status)

Frame details

After first frame is captured, the frame details are available from the tof.Frame() object 

 frameDataDetails = tof.FrameDataDetails()
 status = frames[i].getDataDetails("depth", frameDataDetails)
 print("frame.getDataDetails()", status)
 print("depth frame details:", "width:", frameDataDetails.width, "height:", frameDataDetails.height, "type:", frameDataDetails.type)

Stop Cameras

 for camera in cameras:
    status = camera.stop()
    print("camera"+ IP_addr[cameras.index(camera)]+".stop()", status)
resources/eval/user-guides/eval-adsd3100-nxz-development.txt · Last modified: 05 Dec 2022 05:28 by Harshvardhan Bhatia

Page Tools