Detection of parking slots occupation by temporal difference of inverse perspective mapping from vehicle-borne monocular camera

Detection of parking slots occupation is a crucial task for parking assistance, automatic parking, and autonomous driving systems. This paper proposed a novel method, called Temporal Difference of Inverse Perspective Mapping Difference (TD-IPM), without explicit 3D reconstruction or objection detection. In this method, temporal images from monocular camera are first inverse perspective mapped (IPM) onto the ground plane based on camera calibration results. Second, we proposed an algorithm, called Block Consensus based on Rotation Invariance Phase-Only Correlation (BC-RIPOC), for fast and robust motion estimation. From the estimated motion, we can align these two IPM images and generate IPM difference map. Third, the IPM difference map is segmented and filtered to generate a binary map that can distinguish objects on the ground plane or not for occupation detection. The obstacle is readily localized from the difference map as well. The proposed TD-IPM method has been validated in both underground and outdoor parking lots. Experimental results demonstrate that the proposed TD-IPM method can successfully detect various occupation objects, such as vehicles, cones, lockers, and others, with 97.9% average detection accuracy and speed of 17.5 frames per second (fps). The proposed method suggests an effective and low-cost solution to intelligent parking systems.

Efficient Vehicle Detection and Distance Estimation Based on Aggregated Channel Features and Inverse Perspective Mapping from a Single Camera

In this paper a method for detecting and estimating the distance of a vehicle driving in front using a single black-box camera installed in a vehicle was proposed. In order to apply the proposed method to autonomous vehicles, it was required to reduce the throughput and speed-up the processing. To do this, the proposed method decomposed the input image into multiple-resolution images for real-time processing and then extracted the aggregated channel features (ACFs). The idea was to extract only the most important features from images at different resolutions symmetrically. A method of detecting an object and a method of estimating a vehicle’s distance from a bird’s eye view through inverse perspective mapping (IPM) were applied. In the proposed method, ACFs were used to generate the AdaBoost-based vehicle detector. The ACFs were extracted from the LUV color, edge gradient, and orientation (histograms of oriented gradients) of the input image. Subsequently, by applying IPM and transforming a 2D input image into 3D by generating an image projected in three dimensions, the distance between the detected vehicle and the autonomous vehicle was detected. The proposed method was applied in a real-world road environment and showed accurate results for vehicle detection and distance estimation in real-time processing. Thus, it was showed that our method is applicable to autonomous vehicles.

Computer Vision Measurement of Pointer Meter Readings Based on Inverse Perspective Mapping

A computer vision method for measuring multiple pointer meters is proposed based on the inverse perspective mapping. First, the measured meter scales are used as the calibration objects to obtain the extrinsic parameters of the meter plane. Second, normal vector of the meter plane can be acquired by the extrinsic parameters, obtaining the rotation transformation matrix of the meter image. Then, the acquired meter image is transformed to a position both parallel to the meter plane and near the main point by the rotation transformation matrix and the extrinsic parameters, eliminating the perspective effect of the acquired image. Finally, the transformed image is tested by the visual detection method to obtain the readings of the pointer meter, improving measurement precision. The results of the measurement verify the effectiveness and accuracy of the method.