Vision Integration

Integration Patterns

Vision processing needs to work seamlessly with your robot's control architecture. Here are proven patterns for integrating vision into your codebase.

Pattern 1: Direct OpMode Usage

The simplest approach: use vision detection directly in your autonomous OpMode.

When to use:

Simple autonomous routines
Learning and prototyping
Single-purpose detection

Java Example:

@Autonomous(name = "Simple Vision Auto")
public class SimpleVisionAuto extends LinearOpMode {
    
    private VisionPortal visionPortal;
    private ColorBlobLocatorProcessor colorProcessor;
    
    @Override
    public void runOpMode() {
        // Initialize vision
        colorProcessor = new ColorBlobLocatorProcessor.Builder()
            .setTargetColorRange(ColorRange.BLUE)
            .build();
        
        visionPortal = new VisionPortal.Builder()
            .setCamera(hardwareMap.get(WebcamName.class, "Webcam 1"))
            .addProcessor(colorProcessor)
            .build();
        
        waitForStart();
        
        // Use vision data directly
        List<ColorBlobLocatorProcessor.Blob> blobs = colorProcessor.getBlobs();
        
        if (!blobs.isEmpty()) {
            // Drive to detected object
            double targetX = blobs.get(0).getBoxFit().center.x;
            alignToTarget(targetX);
        }
        
        visionPortal.close();
    }
    
    private void alignToTarget(double targetX) {
        // Alignment logic
    }
}

Pattern 2: Vision Subsystem

Encapsulate vision in a subsystem for cleaner architecture.

When to use:

Multiple OpModes need vision
Complex robot architecture
Team using command-based patterns

Java Example:

package org.firstinspires.ftc.teamcode.subsystems;

import org.firstinspires.ftc.robotcore.external.hardware.camera.WebcamName;
import org.firstinspires.ftc.vision.VisionPortal;
import org.firstinspires.ftc.vision.opencv.ColorBlobLocatorProcessor;
import org.firstinspires.ftc.vision.opencv.ColorRange;
import java.util.List;

public class VisionSubsystem {
    
    private final VisionPortal visionPortal;
    private final ColorBlobLocatorProcessor colorProcessor;
    
    public enum PropPosition {
        LEFT, CENTER, RIGHT, NOT_FOUND
    }
    
    public VisionSubsystem(WebcamName webcamName) {
        // Initialize color processor
        colorProcessor = new ColorBlobLocatorProcessor.Builder()
            .setTargetColorRange(ColorRange.BLUE)
            .setDrawContours(true)
            .setBlurSize(5)
            .build();
        
        // Initialize vision portal
        visionPortal = new VisionPortal.Builder()
            .setCamera(webcamName)
            .addProcessor(colorProcessor)
            .build();
    }
    
    public PropPosition detectPropPosition() {
        List<ColorBlobLocatorProcessor.Blob> blobs = colorProcessor.getBlobs();
        
        // Filter and sort blobs
        ColorBlobLocatorProcessor.Util.filterByArea(500, 10000, blobs);
        ColorBlobLocatorProcessor.Util.sortByArea(
            ColorBlobLocatorProcessor.Util.SortOrder.DESCENDING, blobs
        );
        
        if (blobs.isEmpty()) {
            return PropPosition.NOT_FOUND;
        }
        
        // Determine position
        double centerX = blobs.get(0).getBoxFit().center.x;
        
        if (centerX < 213) return PropPosition.LEFT;
        else if (centerX < 427) return PropPosition.CENTER;
        else return PropPosition.RIGHT;
    }
    
    public List<ColorBlobLocatorProcessor.Blob> getBlobs() {
        return colorProcessor.getBlobs();
    }
    
    public void enableProcessor(boolean enable) {
        visionPortal.setProcessorEnabled(colorProcessor, enable);
    }
    
    public void close() {
        visionPortal.close();
    }
}

Using the Subsystem:

@Autonomous(name = "Vision Subsystem Auto")
public class VisionSubsystemAuto extends LinearOpMode {
    
    private VisionSubsystem vision;
    
    @Override
    public void runOpMode() {
        // Initialize subsystem
        vision = new VisionSubsystem(
            hardwareMap.get(WebcamName.class, "Webcam 1")
        );
        
        // Detect during init
        while (!isStarted()) {
            VisionSubsystem.PropPosition position = vision.detectPropPosition();
            telemetry.addData("Detected Position", position);
            telemetry.update();
        }
        
        // Get final reading
        VisionSubsystem.PropPosition finalPosition = vision.detectPropPosition();
        
        // Disable to save resources
        vision.enableProcessor(false);
        
        waitForStart();
        
        // Execute path based on detection
        switch (finalPosition) {
            case LEFT:
                executeLeftPath();
                break;
            case CENTER:
                executeCenterPath();
                break;
            case RIGHT:
                executeRightPath();
                break;
        }
        
        vision.close();
    }
}

Pattern 3: Mercurial/NextFTC Integration

Integrate with modern command-based frameworks like Mercurial.

Kotlin Example:

package org.firstinspires.ftc.teamcode.subsystems

import com.rowanmcalpin.nextftc.ftc.OpModeData
import com.rowanmcalpin.nextftc.subsystems.Subsystem
import org.firstinspires.ftc.robotcore.external.hardware.camera.WebcamName
import org.firstinspires.ftc.vision.VisionPortal
import org.firstinspires.ftc.vision.opencv.ColorBlobLocatorProcessor
import org.firstinspires.ftc.vision.opencv.ColorRange
import org.firstinspires.ftc.vision.opencv.ImageRegion

class VisionSubsystem : Subsystem {
    
    enum class PropPosition {
        LEFT, CENTER, RIGHT, NOT_FOUND
    }
    
    private val colorProcessor: ColorBlobLocatorProcessor
    private val visionPortal: VisionPortal
    
    var currentPosition: PropPosition = PropPosition.NOT_FOUND
        private set
    
    init {
        // Initialize processor
        colorProcessor = ColorBlobLocatorProcessor.Builder()
            .setTargetColorRange(ColorRange.BLUE)
            .setContourMode(ColorBlobLocatorProcessor.ContourMode.EXTERNAL_ONLY)
            .setRoi(ImageRegion.entireFrame())
            .setDrawContours(true)
            .setBlurSize(5)
            .build()
        
        // Initialize portal
        val webcamName = OpModeData.hardwareMap.get(WebcamName::class.java, "Webcam 1")
        
        visionPortal = VisionPortal.Builder()
            .setCamera(webcamName)
            .addProcessor(colorProcessor)
            .build()
    }
    
    override fun periodic() {
        // Update detection every cycle
        currentPosition = detectPosition()
    }
    
    private fun detectPosition(): PropPosition {
        val blobs = colorProcessor.blobs
        
        // Filter and sort
        ColorBlobLocatorProcessor.Util.filterByArea(500.0, 10000.0, blobs)
        ColorBlobLocatorProcessor.Util.sortByArea(
            ColorBlobLocatorProcessor.Util.SortOrder.DESCENDING, blobs
        )
        
        if (blobs.isEmpty()) {
            return PropPosition.NOT_FOUND
        }
        
        val centerX = blobs[0].boxFit.center.x
        
        return when {
            centerX < 213 -> PropPosition.LEFT
            centerX < 427 -> PropPosition.CENTER
            else -> PropPosition.RIGHT
        }
    }
    
    fun getBlobs() = colorProcessor.blobs
    
    fun enableProcessor(enable: Boolean) {
        visionPortal.setProcessorEnabled(colorProcessor, enable)
    }
    
    fun close() {
        visionPortal.close()
    }
}

Using with Commands:

@Autonomous(name = "Command-Based Vision")
class CommandVisionAuto : OpMode() {
    
    private lateinit var vision: VisionSubsystem
    
    override fun init() {
        OpModeData.updateData(this, gamepad1, gamepad2)
        
        vision = VisionSubsystem()
    }
    
    override fun init_loop() {
        // Vision subsystem automatically updates via periodic()
        telemetry.addData("Detected Position", vision.currentPosition)
        telemetry.update()
    }
    
    override fun start() {
        val detectedPosition = vision.currentPosition
        
        // Build command based on detection
        val autoCommand = when (detectedPosition) {
            VisionSubsystem.PropPosition.LEFT -> buildLeftPath()
            VisionSubsystem.PropPosition.CENTER -> buildCenterPath()
            VisionSubsystem.PropPosition.RIGHT -> buildRightPath()
            VisionSubsystem.PropPosition.NOT_FOUND -> buildDefaultPath()
        }
        
        // Disable vision to save resources
        vision.enableProcessor(false)
        
        // Schedule the command
        CommandScheduler.scheduleCommand(autoCommand)
    }
    
    override fun loop() {
        CommandScheduler.run()
    }
}

Pattern 4: Continuous Tracking

Continuously update robot state using vision data.

Use cases:

AprilTag localization
Dynamic target tracking
Visual servoing

Example:

public class ContinuousVisionTracking extends LinearOpMode {
    
    private AprilTagProcessor aprilTagProcessor;
    private VisionPortal visionPortal;
    private Pose2d robotPose = new Pose2d(0, 0, 0);
    
    @Override
    public void runOpMode() {
        initializeVision();
        
        waitForStart();
        
        // Continuously update pose
        while (opModeIsActive()) {
            updatePoseFromAprilTags();
            
            // Use updated pose for navigation
            navigateToTarget();
            
            telemetry.addData("Robot X", robotPose.getX());
            telemetry.addData("Robot Y", robotPose.getY());
            telemetry.addData("Robot Heading", robotPose.getHeading());
            telemetry.update();
        }
    }
    
    private void updatePoseFromAprilTags() {
        List<AprilTagDetection> detections = aprilTagProcessor.getDetections();
        
        if (!detections.isEmpty()) {
            // Use first visible tag to update pose
            AprilTagDetection detection = detections.get(0);
            
            // Get tag's known field position
            Pose2d tagPose = getTagFieldPose(detection.id);
            
            // Calculate robot pose from tag detection
            robotPose = calculateRobotPose(tagPose, detection);
        }
    }
    
    private Pose2d getTagFieldPose(int tagId) {
        // Return known field position of tag
        // This data comes from field specifications
        return new Pose2d(0, 0, 0);  // Placeholder
    }
    
    private Pose2d calculateRobotPose(Pose2d tagPose, AprilTagDetection detection) {
        // Transform tag-relative pose to field-relative pose
        // Complex math involving rotation matrices
        return new Pose2d(0, 0, 0);  // Placeholder
    }
}

FTC Dashboard Integration

Stream camera feed and telemetry together:

Setup:

Add FTC Dashboard dependency to build.gradle:

implementation 'com.acmerobotics.dashboard:dashboard:0.4.12'

Enable camera streaming:

FtcDashboard dashboard = FtcDashboard.getInstance();
visionPortal = new VisionPortal.Builder()
    .setCamera(hardwareMap.get(WebcamName.class, "Webcam 1"))
    .addProcessor(colorProcessor)
    .build();

// Dashboard automatically shows camera feed

View at http://192.168.43.1:8080/dash

Tunable Parameters:

public class TunableVisionPipeline extends OpenCvPipeline {
    
    @Config
    public static class VisionConfig {
        public static int minHue = 100;
        public static int maxHue = 130;
        public static int minSat = 100;
        public static double minArea = 500;
    }
    
    @Override
    public Mat processFrame(Mat input) {
        // Use VisionConfig values
        Scalar lower = new Scalar(VisionConfig.minHue, VisionConfig.minSat, 100);
        Scalar upper = new Scalar(VisionConfig.maxHue, 255, 255);
        
        // Process with tunable parameters
        return input;
    }
}

Adjust parameters in Dashboard web interface in real-time!

Error Handling

Always handle vision failures gracefully:

public PropPosition detectWithFallback() {
    try {
        List<ColorBlobLocatorProcessor.Blob> blobs = colorProcessor.getBlobs();
        
        if (blobs == null || blobs.isEmpty()) {
            telemetry.addLine("Warning: No blobs detected, using default");
            return PropPosition.CENTER;  // Safe default
        }
        
        // Normal detection logic
        return detectPosition(blobs);
        
    } catch (Exception e) {
        telemetry.addData("Vision Error", e.getMessage());
        telemetry.addLine("Using default position");
        return PropPosition.CENTER;  // Safe fallback
    }
}

Performance Optimization

Lazy Initialization

Only create vision when needed:

private VisionPortal visionPortal;

private void ensureVisionInitialized() {
    if (visionPortal == null) {
        visionPortal = new VisionPortal.Builder()
            .setCamera(hardwareMap.get(WebcamName.class, "Webcam 1"))
            .addProcessor(colorProcessor)
            .build();
    }
}

Conditional Processing

Process vision only when needed:

public class ConditionalVisionSubsystem {
    
    private boolean visionEnabled = false;
    
    public void enableVision() {
        visionEnabled = true;
        visionPortal.setProcessorEnabled(colorProcessor, true);
    }
    
    public void disableVision() {
        visionEnabled = false;
        visionPortal.setProcessorEnabled(colorProcessor, false);
    }
    
    public void periodic() {
        if (visionEnabled) {
            // Only process when enabled
            updateDetections();
        }
    }
}

Testing Strategies

Mock Vision for Testing

public interface IVisionSubsystem {
    PropPosition detectPosition();
}

public class RealVisionSubsystem implements IVisionSubsystem {
    // Real implementation
}

public class MockVisionSubsystem implements IVisionSubsystem {
    private PropPosition mockPosition = PropPosition.CENTER;
    
    public void setMockPosition(PropPosition position) {
        this.mockPosition = position;
    }
    
    @Override
    public PropPosition detectPosition() {
        return mockPosition;
    }
}

Use mock for testing autonomous logic without camera!

Best Practices

Initialize Early - Start camera in init(), not start()
Disable When Done - Save CPU by disabling unused processors
Handle Failures - Always have fallback behavior
Test Thoroughly - Test with and without objects detected
Monitor Performance - Watch for frame drops and lag
Clean Up - Always close VisionPortal when done

Next Steps

Explore Advanced Techniques
See Troubleshooting Guide
Review Migration Guide for upgrading

Integration Patterns

Vision processing needs to work seamlessly with your robot's control architecture. Here are proven patterns for integrating vision into your codebase.

Pattern 1: Direct OpMode Usage

The simplest approach: use vision detection directly in your autonomous OpMode.

When to use:

Simple autonomous routines
Learning and prototyping
Single-purpose detection

Java Example:

@Autonomous(name = "Simple Vision Auto")
public class SimpleVisionAuto extends LinearOpMode {
    
    private VisionPortal visionPortal;
    private ColorBlobLocatorProcessor colorProcessor;
    
    @Override
    public void runOpMode() {
        // Initialize vision
        colorProcessor = new ColorBlobLocatorProcessor.Builder()
            .setTargetColorRange(ColorRange.BLUE)
            .build();
        
        visionPortal = new VisionPortal.Builder()
            .setCamera(hardwareMap.get(WebcamName.class, "Webcam 1"))
            .addProcessor(colorProcessor)
            .build();
        
        waitForStart();
        
        // Use vision data directly
        List<ColorBlobLocatorProcessor.Blob> blobs = colorProcessor.getBlobs();
        
        if (!blobs.isEmpty()) {
            // Drive to detected object
            double targetX = blobs.get(0).getBoxFit().center.x;
            alignToTarget(targetX);
        }
        
        visionPortal.close();
    }
    
    private void alignToTarget(double targetX) {
        // Alignment logic
    }
}

Pattern 2: Vision Subsystem

Encapsulate vision in a subsystem for cleaner architecture.

When to use:

Multiple OpModes need vision
Complex robot architecture
Team using command-based patterns

Java Example:

package org.firstinspires.ftc.teamcode.subsystems;

import org.firstinspires.ftc.robotcore.external.hardware.camera.WebcamName;
import org.firstinspires.ftc.vision.VisionPortal;
import org.firstinspires.ftc.vision.opencv.ColorBlobLocatorProcessor;
import org.firstinspires.ftc.vision.opencv.ColorRange;
import java.util.List;

public class VisionSubsystem {
    
    private final VisionPortal visionPortal;
    private final ColorBlobLocatorProcessor colorProcessor;
    
    public enum PropPosition {
        LEFT, CENTER, RIGHT, NOT_FOUND
    }
    
    public VisionSubsystem(WebcamName webcamName) {
        // Initialize color processor
        colorProcessor = new ColorBlobLocatorProcessor.Builder()
            .setTargetColorRange(ColorRange.BLUE)
            .setDrawContours(true)
            .setBlurSize(5)
            .build();
        
        // Initialize vision portal
        visionPortal = new VisionPortal.Builder()
            .setCamera(webcamName)
            .addProcessor(colorProcessor)
            .build();
    }
    
    public PropPosition detectPropPosition() {
        List<ColorBlobLocatorProcessor.Blob> blobs = colorProcessor.getBlobs();
        
        // Filter and sort blobs
        ColorBlobLocatorProcessor.Util.filterByArea(500, 10000, blobs);
        ColorBlobLocatorProcessor.Util.sortByArea(
            ColorBlobLocatorProcessor.Util.SortOrder.DESCENDING, blobs
        );
        
        if (blobs.isEmpty()) {
            return PropPosition.NOT_FOUND;
        }
        
        // Determine position
        double centerX = blobs.get(0).getBoxFit().center.x;
        
        if (centerX < 213) return PropPosition.LEFT;
        else if (centerX < 427) return PropPosition.CENTER;
        else return PropPosition.RIGHT;
    }
    
    public List<ColorBlobLocatorProcessor.Blob> getBlobs() {
        return colorProcessor.getBlobs();
    }
    
    public void enableProcessor(boolean enable) {
        visionPortal.setProcessorEnabled(colorProcessor, enable);
    }
    
    public void close() {
        visionPortal.close();
    }
}

Using the Subsystem:

@Autonomous(name = "Vision Subsystem Auto")
public class VisionSubsystemAuto extends LinearOpMode {
    
    private VisionSubsystem vision;
    
    @Override
    public void runOpMode() {
        // Initialize subsystem
        vision = new VisionSubsystem(
            hardwareMap.get(WebcamName.class, "Webcam 1")
        );
        
        // Detect during init
        while (!isStarted()) {
            VisionSubsystem.PropPosition position = vision.detectPropPosition();
            telemetry.addData("Detected Position", position);
            telemetry.update();
        }
        
        // Get final reading
        VisionSubsystem.PropPosition finalPosition = vision.detectPropPosition();
        
        // Disable to save resources
        vision.enableProcessor(false);
        
        waitForStart();
        
        // Execute path based on detection
        switch (finalPosition) {
            case LEFT:
                executeLeftPath();
                break;
            case CENTER:
                executeCenterPath();
                break;
            case RIGHT:
                executeRightPath();
                break;
        }
        
        vision.close();
    }
}

Pattern 3: Mercurial/NextFTC Integration

Integrate with modern command-based frameworks like Mercurial.

Kotlin Example:

package org.firstinspires.ftc.teamcode.subsystems

import com.rowanmcalpin.nextftc.ftc.OpModeData
import com.rowanmcalpin.nextftc.subsystems.Subsystem
import org.firstinspires.ftc.robotcore.external.hardware.camera.WebcamName
import org.firstinspires.ftc.vision.VisionPortal
import org.firstinspires.ftc.vision.opencv.ColorBlobLocatorProcessor
import org.firstinspires.ftc.vision.opencv.ColorRange
import org.firstinspires.ftc.vision.opencv.ImageRegion

class VisionSubsystem : Subsystem {
    
    enum class PropPosition {
        LEFT, CENTER, RIGHT, NOT_FOUND
    }
    
    private val colorProcessor: ColorBlobLocatorProcessor
    private val visionPortal: VisionPortal
    
    var currentPosition: PropPosition = PropPosition.NOT_FOUND
        private set
    
    init {
        // Initialize processor
        colorProcessor = ColorBlobLocatorProcessor.Builder()
            .setTargetColorRange(ColorRange.BLUE)
            .setContourMode(ColorBlobLocatorProcessor.ContourMode.EXTERNAL_ONLY)
            .setRoi(ImageRegion.entireFrame())
            .setDrawContours(true)
            .setBlurSize(5)
            .build()
        
        // Initialize portal
        val webcamName = OpModeData.hardwareMap.get(WebcamName::class.java, "Webcam 1")
        
        visionPortal = VisionPortal.Builder()
            .setCamera(webcamName)
            .addProcessor(colorProcessor)
            .build()
    }
    
    override fun periodic() {
        // Update detection every cycle
        currentPosition = detectPosition()
    }
    
    private fun detectPosition(): PropPosition {
        val blobs = colorProcessor.blobs
        
        // Filter and sort
        ColorBlobLocatorProcessor.Util.filterByArea(500.0, 10000.0, blobs)
        ColorBlobLocatorProcessor.Util.sortByArea(
            ColorBlobLocatorProcessor.Util.SortOrder.DESCENDING, blobs
        )
        
        if (blobs.isEmpty()) {
            return PropPosition.NOT_FOUND
        }
        
        val centerX = blobs[0].boxFit.center.x
        
        return when {
            centerX < 213 -> PropPosition.LEFT
            centerX < 427 -> PropPosition.CENTER
            else -> PropPosition.RIGHT
        }
    }
    
    fun getBlobs() = colorProcessor.blobs
    
    fun enableProcessor(enable: Boolean) {
        visionPortal.setProcessorEnabled(colorProcessor, enable)
    }
    
    fun close() {
        visionPortal.close()
    }
}

Using with Commands:

@Autonomous(name = "Command-Based Vision")
class CommandVisionAuto : OpMode() {
    
    private lateinit var vision: VisionSubsystem
    
    override fun init() {
        OpModeData.updateData(this, gamepad1, gamepad2)
        
        vision = VisionSubsystem()
    }
    
    override fun init_loop() {
        // Vision subsystem automatically updates via periodic()
        telemetry.addData("Detected Position", vision.currentPosition)
        telemetry.update()
    }
    
    override fun start() {
        val detectedPosition = vision.currentPosition
        
        // Build command based on detection
        val autoCommand = when (detectedPosition) {
            VisionSubsystem.PropPosition.LEFT -> buildLeftPath()
            VisionSubsystem.PropPosition.CENTER -> buildCenterPath()
            VisionSubsystem.PropPosition.RIGHT -> buildRightPath()
            VisionSubsystem.PropPosition.NOT_FOUND -> buildDefaultPath()
        }
        
        // Disable vision to save resources
        vision.enableProcessor(false)
        
        // Schedule the command
        CommandScheduler.scheduleCommand(autoCommand)
    }
    
    override fun loop() {
        CommandScheduler.run()
    }
}

Pattern 4: Continuous Tracking

Continuously update robot state using vision data.

Use cases:

AprilTag localization
Dynamic target tracking
Visual servoing

Example:

public class ContinuousVisionTracking extends LinearOpMode {
    
    private AprilTagProcessor aprilTagProcessor;
    private VisionPortal visionPortal;
    private Pose2d robotPose = new Pose2d(0, 0, 0);
    
    @Override
    public void runOpMode() {
        initializeVision();
        
        waitForStart();
        
        // Continuously update pose
        while (opModeIsActive()) {
            updatePoseFromAprilTags();
            
            // Use updated pose for navigation
            navigateToTarget();
            
            telemetry.addData("Robot X", robotPose.getX());
            telemetry.addData("Robot Y", robotPose.getY());
            telemetry.addData("Robot Heading", robotPose.getHeading());
            telemetry.update();
        }
    }
    
    private void updatePoseFromAprilTags() {
        List<AprilTagDetection> detections = aprilTagProcessor.getDetections();
        
        if (!detections.isEmpty()) {
            // Use first visible tag to update pose
            AprilTagDetection detection = detections.get(0);
            
            // Get tag's known field position
            Pose2d tagPose = getTagFieldPose(detection.id);
            
            // Calculate robot pose from tag detection
            robotPose = calculateRobotPose(tagPose, detection);
        }
    }
    
    private Pose2d getTagFieldPose(int tagId) {
        // Return known field position of tag
        // This data comes from field specifications
        return new Pose2d(0, 0, 0);  // Placeholder
    }
    
    private Pose2d calculateRobotPose(Pose2d tagPose, AprilTagDetection detection) {
        // Transform tag-relative pose to field-relative pose
        // Complex math involving rotation matrices
        return new Pose2d(0, 0, 0);  // Placeholder
    }
}

FTC Dashboard Integration

Stream camera feed and telemetry together:

Setup:

Add FTC Dashboard dependency to build.gradle:

implementation 'com.acmerobotics.dashboard:dashboard:0.4.12'

Enable camera streaming:

FtcDashboard dashboard = FtcDashboard.getInstance();
visionPortal = new VisionPortal.Builder()
    .setCamera(hardwareMap.get(WebcamName.class, "Webcam 1"))
    .addProcessor(colorProcessor)
    .build();

// Dashboard automatically shows camera feed

View at http://192.168.43.1:8080/dash

Tunable Parameters:

public class TunableVisionPipeline extends OpenCvPipeline {
    
    @Config
    public static class VisionConfig {
        public static int minHue = 100;
        public static int maxHue = 130;
        public static int minSat = 100;
        public static double minArea = 500;
    }
    
    @Override
    public Mat processFrame(Mat input) {
        // Use VisionConfig values
        Scalar lower = new Scalar(VisionConfig.minHue, VisionConfig.minSat, 100);
        Scalar upper = new Scalar(VisionConfig.maxHue, 255, 255);
        
        // Process with tunable parameters
        return input;
    }
}

Adjust parameters in Dashboard web interface in real-time!

Error Handling

Always handle vision failures gracefully:

public PropPosition detectWithFallback() {
    try {
        List<ColorBlobLocatorProcessor.Blob> blobs = colorProcessor.getBlobs();
        
        if (blobs == null || blobs.isEmpty()) {
            telemetry.addLine("Warning: No blobs detected, using default");
            return PropPosition.CENTER;  // Safe default
        }
        
        // Normal detection logic
        return detectPosition(blobs);
        
    } catch (Exception e) {
        telemetry.addData("Vision Error", e.getMessage());
        telemetry.addLine("Using default position");
        return PropPosition.CENTER;  // Safe fallback
    }
}

Performance Optimization

Lazy Initialization

Only create vision when needed:

private VisionPortal visionPortal;

private void ensureVisionInitialized() {
    if (visionPortal == null) {
        visionPortal = new VisionPortal.Builder()
            .setCamera(hardwareMap.get(WebcamName.class, "Webcam 1"))
            .addProcessor(colorProcessor)
            .build();
    }
}

Conditional Processing

Process vision only when needed:

public class ConditionalVisionSubsystem {
    
    private boolean visionEnabled = false;
    
    public void enableVision() {
        visionEnabled = true;
        visionPortal.setProcessorEnabled(colorProcessor, true);
    }
    
    public void disableVision() {
        visionEnabled = false;
        visionPortal.setProcessorEnabled(colorProcessor, false);
    }
    
    public void periodic() {
        if (visionEnabled) {
            // Only process when enabled
            updateDetections();
        }
    }
}

Testing Strategies

Mock Vision for Testing

public interface IVisionSubsystem {
    PropPosition detectPosition();
}

public class RealVisionSubsystem implements IVisionSubsystem {
    // Real implementation
}

public class MockVisionSubsystem implements IVisionSubsystem {
    private PropPosition mockPosition = PropPosition.CENTER;
    
    public void setMockPosition(PropPosition position) {
        this.mockPosition = position;
    }
    
    @Override
    public PropPosition detectPosition() {
        return mockPosition;
    }
}

Use mock for testing autonomous logic without camera!

Best Practices

Initialize Early - Start camera in init(), not start()
Disable When Done - Save CPU by disabling unused processors
Handle Failures - Always have fallback behavior
Test Thoroughly - Test with and without objects detected
Monitor Performance - Watch for frame drops and lag
Clean Up - Always close VisionPortal when done

Next Steps

Explore Advanced Techniques
See Troubleshooting Guide
Review Migration Guide for upgrading

Vision Integration

Integration Patterns

Pattern 1: Direct OpMode Usage

Pattern 2: Vision Subsystem

Pattern 3: Mercurial/NextFTC Integration

Pattern 4: Continuous Tracking

FTC Dashboard Integration

Error Handling

Performance Optimization

Lazy Initialization

Conditional Processing

Testing Strategies

Mock Vision for Testing

Best Practices

Next Steps

On this page

Vision Integration

Integration Patterns

Pattern 1: Direct OpMode Usage

Pattern 2: Vision Subsystem

Pattern 3: Mercurial/NextFTC Integration

Pattern 4: Continuous Tracking

FTC Dashboard Integration

Error Handling

Performance Optimization

Lazy Initialization

Conditional Processing

Testing Strategies

Mock Vision for Testing

Best Practices

Next Steps

On this page