Scene Recognition TRIZ Cases
- Radar-Camera Object Recognition System for Large Machines
Processor maps radar detections to image regions for targeted object classification, reducing computational load on large machines.
- Covariance Appearance Models for Object Detection
Segmented covariance appearance models detect objects while reducing computational intensity.
- In-path Object Detection Using Inertial and Radar Data
A detection system uses inertial measurement unit data to determine lane curvature and boundaries without image processing.
- Automated Metric Information Network for Orthomosaic Alignment
Automated metric information network reduces orthomosaic misalignment errors by generating ground control points through automated tie pointing and clustering.
- Land Cover Classification Using Temporal Sequences and Viterbi Algorithm
Aerial image classifier uses temporal sequences to identify land cover states via a doubly embedded stochastic process.
- Vehicle LiDAR Road Boundary Detection Grid Segmentation
A vehicle LiDAR system segments detection space into lane grids to calculate road boundary candidates using occupation percentages and freespace point data.
- Trailer Edge Tracking via Geometric Modeling
Geometric modeling replaces complex motion detection to maintain stable framing during cornering without excessive computational demand.
- Vehicle Target Confidence Scoring via Sensor Weighting
A computer assigns weighting coefficients to detected targets based on sensor characteristics and environmental conditions.
- Onboard Camera Data Extraction and Classification System
A data processing system extracts and classifies feature information from onboard camera images.
- Vehicle Drive-Assist Apparatus Urban Acceleration Control
A vehicle drive-assist apparatus suppresses engine output in urban regions to prevent unintended acceleration from pedal errors.
- Oncoming Vehicle Detection for Waste Collection Safety
Optical sensors detect approaching vehicles and trigger alarms, reducing collision injury risk during waste collection operations.
- 3D Object Annotation via 2D Boundary Projection
A system projects LiDAR point clouds onto camera images to detect object boundaries.
- CNN Object and Lane Detection System
A convolutional neural network system processes input images to detect objects and lane markings simultaneously using shared feature extraction layers.
- Autonomous Driving Control Apparatus Using Moving Object Removal
A vehicle control apparatus detects moving objects in real-time images to isolate static background features for accurate position recognition.
- Deep Learning Machine Adaptive Integer Quantization
A deep learning machine classifies input data and applies adaptive integer quantization for efficient inference.
- Vehicle Sensor False Negative Rate Assessment by Time Class
Classifies environmental sensor false negatives into time classes to determine performance impairment thresholds for automated driving systems.