Processor maps radar detections to image regions for targeted object classification, reducing computational load on large machines.
Segmented covariance appearance models detect objects while reducing computational intensity.
A detection system uses inertial measurement unit data to determine lane curvature and boundaries without image processing.
Automated metric information network reduces orthomosaic misalignment errors by generating ground control points through automated tie pointing and clustering.
Aerial image classifier uses temporal sequences to identify land cover states via a doubly embedded stochastic process.
A vehicle LiDAR system segments detection space into lane grids to calculate road boundary candidates using occupation percentages and freespace point data.
Geometric modeling replaces complex motion detection to maintain stable framing during cornering without excessive computational demand.
A computer assigns weighting coefficients to detected targets based on sensor characteristics and environmental conditions.
A data processing system extracts and classifies feature information from onboard camera images.
A vehicle drive-assist apparatus suppresses engine output in urban regions to prevent unintended acceleration from pedal errors.
Optical sensors detect approaching vehicles and trigger alarms, reducing collision injury risk during waste collection operations.
A system projects LiDAR point clouds onto camera images to detect object boundaries.
A convolutional neural network system processes input images to detect objects and lane markings simultaneously using shared feature extraction layers.
A vehicle control apparatus detects moving objects in real-time images to isolate static background features for accurate position recognition.
A deep learning machine classifies input data and applies adaptive integer quantization for efficient inference.
Classifies environmental sensor false negatives into time classes to determine performance impairment thresholds for automated driving systems.