4 results about "Picture plane" patented technology

A method for generating an image of a three-dimensional object in a virtual three-dimensional space includes using a virtual camera in the three-dimensional space. The method includes updating a position and / or orientation of the virtual camera in response to a change in a position of at least one vanishing point existing on a picture plane and corresponding to infinity of coordinate axes. The method includes generating the image of the three-dimensional object using the updated virtual camera. The updating includes changing the position and / or orientation of the virtual camera in a case where the changed position of the vanishing point is returned to the position of the vanishing point before the change in the position of the vanishing point, the updated position and / or orientation of the virtual camera is returned to the position and / or orientation of the virtual camera before the change in the position of the vanishing point.

This object recognition method comprises: projecting detected point-cloud objects detected from point cloud data onto an image plane; determining whether the positions of detected image objects and the positions of the detected point-cloud objects are close to each other on the image plane; calculating, for each detected image object and detected point-cloud object positioned close to each other, a shape feature quantity indicating a feature quantity of the shape on the image plane; calculating a degree of similarity on the basis of the shape feature quantities respectively calculated for the detected image object and the detected point-cloud object; and determining that the detected image object and the detected point-cloud object are the same if the degree of similarity is greater than a threshold.

The invention relates to an optical system (28) for providing an image (12) of an object of interest (1, 6), with the image (12) being shiftable (5, 13) in an image plane of the optical system (28) and with the object of interest (1, 6) being located in an object plane of the optical system (28) and with the image plane being conjugate with respect to the object plane. The optical system (28) comprises 1) a first lens module (7, 9, 30) facing the object plane, the first lens module (7, 9, 30) having a first optical axis and a first entrance pupil, which first entrance pupil is located outside the first lens module (7, 9, 30) and is, with respect to the first optical axis, facing the opposite side of the first lens module (7, 9, 30) compared to the side of the first lens module (7, 9, 30) facing the object plane, a second 2) a second optical axis and a second entrance pupil, which second entrance pupil is located outside the second scan lens module (11) and is, with respect to the second optical axis, facing the opposite side of the second scan lens module (11) compared to the side of the second scan lens module facing the image plane, wherein the second scan lens module (11) is arranged such that the first optical axis and the second optical axis are coplanar and orthogonal to one another, and 3) a at least one mirror (10), with each mirror (10) of the at least one mirror (10) being rotatable about a respective axis of rotation and with the at least one mirror (10) being arranged at the first entrance pupil and at the second entrance pupil, and with all axes of rotation of the at least one axis of rotation (i) being parallel to one another, (ii) being coplanar to one another and lying in a rotation plane, and (iii) being orthogonal to the first optical axis and to the second optical axis, wherein the at least one mirror (10) is arranged such that the rotation plane passes through an intersection of (i) the first optical axis and (ii) the second optical axis. The invention also relates to a hyperspectral camera (4) comprising the optical system (28) and an imaging spectrometer (29) and to a method for capturing spectrally resolved data using the hyperspectral camera (4).

The application discloses an image plane region detection method based on edge information. Firstly, a neural network is used to extract basic skeleton network features of each level of an image. Then, each level feature obtained is extracted into multi-level context features through upsampling and interlayer linking. The intermediate level feature obtained is extracted into multi-level edge features of the image through iterative optimization network layers, and different level features are optimized into expected feature dimensions. The multi-level context features and the multi-level edge features are fused in pairs layer by layer. Finally, the fused features are provided to a target model of a downstream image plane detection task to be trained, so as to improve the performance of the target model in the downstream image plane detection task.