API Reference - Self-Driving Vision Module

Object Detection API

detectObjects

POST

Detects objects in an image using YOLO v8 model. Returns bounding boxes, confidence scores, and class labels for detected objects.

Endpoint

POST /api/v1/detect-objects

Parameters

Parameter	Type	Required	Description
image	File	Yes	Image file (JPEG, PNG, BMP)
confidence	float	No	Confidence threshold (0.0-1.0, default: 0.5)
model	string	No	Model variant (yolov8n, yolov8s, yolov8m, yolov8l)
classes	array	No	Filter specific object classes

Example Request

import requests

url = "https://api.selfdriving-vision.com/v1/detect-objects"
files = {"image": open("road_scene.jpg", "rb")}
data = {
    "confidence": 0.6,
    "model": "yolov8m",
    "classes": ["person", "car", "truck", "bus"]
}

response = requests.post(url, files=files, data=data)
                

Response

Success Response (200 OK)

{
    "status": "success",
    "detections": [
        {
            "class": "car",
            "confidence": 0.94,
            "bbox": {
                "x": 245,
                "y": 156,
                "width": 89,
                "height": 67
            },
            "center": {
                "x": 289.5,
                "y": 189.5
            }
        },
        {
            "class": "person",
            "confidence": 0.87,
            "bbox": {
                "x": 123,
                "y": 203,
                "width": 34,
                "height": 78
            },
            "center": {
                "x": 140,
                "y": 242
            }
        }
    ],
    "processing_time": 0.045,
    "model_version": "yolov8m-v1.2.3"
}
                    

getModelInfo

GET

Retrieves information about available detection models and their capabilities.

GET /api/v1/models

Success Response (200 OK)

{
    "models": [
        {
            "name": "yolov8n",
            "description": "YOLO v8 Nano - Fastest inference",
            "classes": ["person", "bicycle", "car", "motorcycle", "airplane", "bus", "train", "truck"],
            "avg_inference_time": 0.012,
            "model_size": "6.2MB",
            "mAP": 0.374
        },
        {
            "name": "yolov8l",
            "description": "YOLO v8 Large - Highest accuracy",
            "classes": ["person", "bicycle", "car", "motorcycle", "airplane", "bus", "train", "truck"],
            "avg_inference_time": 0.089,
            "model_size": "87.7MB",
            "mAP": 0.531
        }
    ]
}
                    

Lane Detection API

detectLanes

POST

Detects lane boundaries in road images using computer vision algorithms. Returns lane lines with curvature and positioning information.

POST /api/v1/detect-lanes

Parameters

Parameter	Type	Required	Description
image	File	Yes	Road image file
algorithm	string	No	Detection algorithm (canny, hough, cnn)
roi	object	No	Region of interest coordinates

Success Response (200 OK)

{
    "status": "success",
    "lanes": [
        {
            "side": "left",
            "points": [
                {"x": 120, "y": 400},
                {"x": 135, "y": 350},
                {"x": 150, "y": 300}
            ],
            "curvature": 0.0023,
            "confidence": 0.89
        },
        {
            "side": "right",
            "points": [
                {"x": 680, "y": 400},
                {"x": 665, "y": 350},
                {"x": 650, "y": 300}
            ],
            "curvature": 0.0018,
            "confidence": 0.92
        }
    ],
    "lane_width": 3.2,
    "vehicle_offset": -0.15,
    "processing_time": 0.034
}
                    

Traffic Sign Recognition API

recognizeTrafficSigns

POST

Recognizes and classifies traffic signs in images using CNN models trained on traffic sign datasets.

POST /api/v1/recognize-traffic-signs

Parameters

Parameter	Type	Required	Description
image	File	Yes	Image containing traffic signs
confidence	float	No	Minimum confidence threshold
categories	array	No	Filter by sign categories

Success Response (200 OK)

{
    "status": "success",
    "signs": [
        {
            "class": "stop",
            "confidence": 0.95,
            "bbox": {
                "x": 245,
                "y": 156,
                "width": 89,
                "height": 89
            },
            "category": "regulatory",
            "description": "Stop sign - Come to a complete stop"
        },
        {
            "class": "speed_limit_50",
            "confidence": 0.88,
            "bbox": {
                "x": 123,
                "y": 203,
                "width": 67,
                "height": 67
            },
            "category": "regulatory",
            "description": "Speed limit 50 km/h"
        }
    ],
    "processing_time": 0.067
}
                    

Pedestrian Detection API

detectPedestrians

POST

Detects pedestrians in images with pose estimation and movement prediction capabilities.

POST /api/v1/detect-pedestrians

Parameters

Parameter	Type	Required	Description
image	File	Yes	Image containing pedestrians
pose_estimation	boolean	No	Enable pose keypoint detection
movement_prediction	boolean	No	Enable trajectory prediction

Success Response (200 OK)

{
    "status": "success",
    "pedestrians": [
        {
            "id": 1,
            "confidence": 0.92,
            "bbox": {
                "x": 123,
                "y": 203,
                "width": 34,
                "height": 78
            },
            "pose": {
                "head": {"x": 140, "y": 210},
                "left_shoulder": {"x": 135, "y": 225},
                "right_shoulder": {"x": 145, "y": 225},
                "left_hip": {"x": 138, "y": 265},
                "right_hip": {"x": 142, "y": 265}
            },
            "movement": {
                "velocity": 1.2,
                "direction": 45.5,
                "predicted_position": {"x": 148, "y": 220}
            }
        }
    ],
    "processing_time": 0.089
}
                    

Sensor Fusion API

fuseSensors

POST

Combines data from multiple sensors (camera, LiDAR, radar) to create a unified perception of the environment.

POST /api/v1/fuse-sensors

Parameters

Parameter	Type	Required	Description
camera_data	File	Yes	Camera image data
lidar_data	File	No	LiDAR point cloud data
radar_data	JSON	No	Radar detection data
fusion_algorithm	string	No	Fusion method (kalman, bayesian, neural)

Success Response (200 OK)

{
    "status": "success",
    "fused_objects": [
        {
            "id": 1,
            "class": "car",
            "confidence": 0.94,
            "position": {
                "x": 15.2,
                "y": -2.1,
                "z": 0.5
            },
            "velocity": {
                "x": 12.5,
                "y": 0.0,
                "z": 0.0
            },
            "dimensions": {
                "length": 4.2,
                "width": 1.8,
                "height": 1.5
            },
            "sensor_contributions": {
                "camera": 0.6,
                "lidar": 0.8,
                "radar": 0.7
            }
        }
    ],
    "processing_time": 0.156,
    "fusion_confidence": 0.89
}
                    

Error Handling

HTTP Status Codes

Code	Description
200	Success
400	Bad Request - Invalid parameters
401	Unauthorized - Invalid API key
413	Payload Too Large - Image too large
429	Too Many Requests - Rate limit exceeded
500	Internal Server Error

Error Response Format

{
    "status": "error",
    "error": {
        "code": "INVALID_IMAGE_FORMAT",
        "message": "Unsupported image format. Please use JPEG, PNG, or BMP.",
        "details": {
            "received_format": "GIF",
            "supported_formats": ["JPEG", "PNG", "BMP"]
        }
    },
    "timestamp": "2024-01-15T10:30:00Z"
}
                