The present disclosure provides "sensor fusion". A plurality of images may be acquired from a plurality of sensors, and a plurality of flattened tiles may be extracted from the plurality of images. An image location in the plurality of images and a sensor type token may be added to each of the plurality of flattened tiles, the sensor type token identifying a sensor type used to acquire an image of the plurality of images from which the respective flattened tile was acquired. The flattened tiles may be joined into a flattened tensor, and a task token indicating a processing task is added to the flattened tensor, where the flattened tensor is a one-dimensional array comprising two or more data types. The flattened tensor may be input to a first deep neural network that includes a plurality of encoder layers and a plurality of decoder layers and outputs a transformer output. The transformer output may be input to a second deep neural network, which determines the object prediction indicated by the token, and may output the object prediction.
本公开提供“传感器融合”。可以从多个传感器获取多个图像,并且可以从多个图像中提取多个扁平化片块。可以向多个扁平化片块中的每一个添加在多个图像中的图像位置以及传感器类型令牌,传感器类型令牌标识用于获取多个图像中的、从中获取了相应扁平化片块的图像的传感器类型。扁平化片块可以被联接成扁平张量,并向扁平张量添加指示处理任务的任务令牌,其中扁平张量是包括两种或更多种数据类型的一维阵列。扁平张量可以被输入到第一深度神经网络,第一深度神经网络包括多个编码器层和多个解码器层并输出变换器输出。可以将变换器输出输入到第二深度神经网络,第二深度神经网络确定由令牌指示的对象预测,并且可以输出对象预测。
Sensor fusion
传感器融合
2023-07-28
Patent
Elektronische Ressource
Chinesisch
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06V / H04N PICTORIAL COMMUNICATION, e.g. TELEVISION , Bildübertragung, z.B. Fernsehen / H04W WIRELESS COMMUNICATION NETWORKS , Drahtlose Kommunikationsnetze |
SENSOR FUSION SYSTEM DRIVING ASSISTANCE SYSTEM AND SENSOR FUSION METHOD
Europäisches Patentamt | 2024
|