CN-121982665-A - Complex scene traffic perception method and system based on fuzzy logic and data enhancement

CN121982665ACN 121982665 ACN121982665 ACN 121982665ACN-121982665-A

Abstract

The invention discloses a complex scene traffic perception method and system based on fuzzy logic and data enhancement, and relates to the technical field of automatic driving environment perception. The method comprises the steps of obtaining RGB images and point cloud data, performing preprocessing operation to obtain a sample set, expanding the sample set by using a forced enhancement mechanism to obtain an expanded sample set, performing exclusive feature enhancement operation on the expanded sample set according to scene complexity to obtain enhanced features, performing dynamic weight calculation and feature fusion on the enhanced features by using a fuzzy logic algorithm based on factors influencing traffic perception to obtain fusion features, and performing perception prediction on the fusion features by using a perception prediction model to obtain a perception prediction result. The method combines the forced weather enhancement and fuzzy logic fusion technology, adapts to complex scenes, and solves the difficult problems of insufficient multi-mode perception robustness and generalization in severe weather.

Inventors

LI YANG
XIE XIAODONG
Yang tianjiao
SUN ZHENBANG
WANG JIYANG
YANG RONGDONG

Assignees

山东省科学院自动化研究所

Dates

Publication Date: 20260505
Application Date: 20251211

Claims (10)

1. The complex scene traffic perception method based on fuzzy logic and data enhancement is characterized by comprising the following steps of: acquiring RGB images and point cloud data, and performing preprocessing operation to obtain a sample set; Expanding the sample set by using a forced enhancement mechanism to obtain an expanded sample set; Carrying out exclusive characteristic enhancement operation on the expanded sample set according to scene complexity to obtain enhanced characteristics; performing dynamic weight calculation and feature fusion on the enhanced features based on factors influencing traffic perception by using a fuzzy logic algorithm to obtain fusion features; and performing perception prediction on the fusion characteristics by using a perception prediction model to obtain a perception prediction result.
2. The complex scene traffic perception method based on fuzzy logic and data enhancement as claimed in claim 1, wherein the specific steps of performing preprocessing operation are: performing image noise and high-frequency interference removal operation on the RGB image; filtering the point cloud data and calculating intensity variance; and labeling the preprocessed RGB image and the point cloud data.
3. The complex scene traffic perception method based on fuzzy logic and data enhancement as claimed in claim 1, wherein the specific steps of expanding the sample set by using the forced enhancement mechanism are as follows: And constructing a forced enhancement mechanism, and respectively carrying out data enhancement processing on three types of weather, namely rain, fog and snow by utilizing the forced enhancement mechanism, specifically adding corresponding weather characteristic interference to an RGB image, and carrying out noise injection or attenuation processing on point cloud data.
4. The complex scene traffic perception method based on fuzzy logic and data enhancement as claimed in claim 1, wherein the specific steps of performing dedicated feature enhancement operation on the extended sample set according to scene complexity are as follows: Determining scene complexity based on the scene description keywords and the target number; The expanded sample set is divided into 4 categories according to scene complexity, namely a simple scene, a congestion scene, a construction scene and an intersection scene; and respectively executing exclusive characteristic enhancement on the congestion scene and the construction scene.
5. The complex scene traffic perception method based on fuzzy logic and data enhancement as claimed in claim 4, wherein the specific steps of performing dedicated feature enhancement on the congestion scene and the construction scene respectively are as follows: congestion scene enhancement is performed by adding a region of interest; and reinforcing the construction scene by reinforcing the static obstacle characteristics.
6. The complex scene traffic perception method based on fuzzy logic and data enhancement as claimed in claim 1, wherein the specific steps of dynamic weight calculation and feature fusion of the enhancement features based on factors affecting traffic perception by using a fuzzy logic algorithm are as follows: Constructing a fuzzy logic framework; weight reasoning is carried out according to the fuzzy logic framework; And carrying out feature fusion according to the weight reasoning result.
7. The complex scene traffic perception method based on fuzzy logic and data enhancement as claimed in claim 6, wherein the specific steps of weight reasoning according to the fuzzy logic framework are as follows: performing blurring on input parameters of all samples in the batch according to the fuzzy logic framework; calculating the triggering strength of each rule according to a minimum membership method, and carrying out weight aggregation according to the triggering strength; The defuzzification is performed using the barycenter method.
8. The complex scene traffic perception system based on fuzzy logic and data enhancement is characterized by comprising the following components: the data acquisition module is configured to acquire RGB images and point cloud data, and perform preprocessing operation to obtain a sample set; the data expansion module is configured to expand the sample set by utilizing a forced enhancement mechanism to obtain an expanded sample set; The characteristic enhancement module is configured to perform exclusive characteristic enhancement operation on the expanded sample set according to scene complexity to obtain enhanced characteristics; The fuzzy logic module is configured to utilize a fuzzy logic algorithm to perform dynamic weight calculation and feature fusion on the enhanced features based on factors influencing traffic perception, so as to obtain fusion features; And the perception prediction module is configured to carry out perception prediction on the fusion characteristics by using a perception prediction model to obtain a perception prediction result.
9. A computer readable storage medium, characterized in that it stores a computer program adapted to be loaded by a processor and to perform the complex scene traffic awareness method based on fuzzy logic and data enhancement according to any of claims 1-7.
10. A computer device, comprising: A processor adapted to execute a computer program; a computer readable storage medium having stored therein a computer program which, when executed by the processor, implements the complex scene traffic awareness method based on fuzzy logic and data enhancement of any of claims 1-7.

Description

Complex scene traffic perception method and system based on fuzzy logic and data enhancement Technical Field The invention relates to the technical field of automatic driving environment sensing, in particular to a complex scene traffic sensing method and system based on fuzzy logic and data enhancement. Background The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art. In recent years, deep learning technology is increasingly widely applied in the field of automatic driving environment awareness. Along with the evolution of automatic driving to the L4 level and higher high-order stages, a multi-mode sensor (a camera provides visual semantic information and a laser radar provides three-dimensional space information) fusion technology is gradually developed into an essential core tool of automatic driving safety driving by means of the detection advantage of 'visual-laser' complementation. Along with the continuous progress of sensor manufacturing technology, particularly the development of high-resolution cameras (such as 8MP forward-looking cameras) and low-cost laser radars (such as 128-line hybrid solid-state radars), a large number of automatic driving vehicle types realize multi-sensor standard allocation, so that the multi-mode fusion perception technology is promoted to move from laboratory simulation to real road scenes, and the transition from 'single sensor dependence' to 'multi-mode cooperation' is realized. The potential of the multi-mode fusion technology in the aspects of sensing precision, environmental adaptability and complex scene coverage capability is continuously mined, and the multi-mode fusion sensing is gradually one of the core technologies of an automatic driving system under the driving of market demands, particularly in diversified driving scenes such as urban roads, highways, construction areas and the like. However, despite significant advances in multi-modal fusion techniques, current algorithms and systems still face a number of challenges, particularly in terms of robustness, generalization capability, and real-time balance. Robustness is limited by severe weather and complex scene disturbances, resulting in model instability. The generalization capability of the existing algorithm is insufficient, and the false detection rate is high. The insufficient number of samples results in serious problems of overfitting and real-time imbalance. Disclosure of Invention Aiming at the defects of the prior art, the invention aims to provide a complex scene traffic perception method and system based on fuzzy logic and data enhancement, which combines forced weather enhancement and fuzzy logic fusion technology and adapts to complex scenes, thereby solving the difficult problems of insufficient multi-mode perception robustness and generalization in severe weather. In order to achieve the above object, the present invention is realized by the following technical scheme: The invention provides a complex scene traffic perception method based on fuzzy logic and data enhancement, which comprises the following steps: acquiring RGB images and point cloud data, and performing preprocessing operation to obtain a sample set; Expanding the sample set by using a forced enhancement mechanism to obtain an expanded sample set; Carrying out exclusive characteristic enhancement operation on the expanded sample set according to scene complexity to obtain enhanced characteristics; performing dynamic weight calculation and feature fusion on the enhanced features based on factors influencing traffic perception by using a fuzzy logic algorithm to obtain fusion features; and performing perception prediction on the fusion characteristics by using a perception prediction model to obtain a perception prediction result. The second aspect of the present invention provides a complex scene traffic perception system based on fuzzy logic and data enhancement, comprising: the data acquisition module is configured to acquire RGB images and point cloud data, and perform preprocessing operation to obtain a sample set; the data expansion module is configured to expand the sample set by utilizing a forced enhancement mechanism to obtain an expanded sample set; The characteristic enhancement module is configured to perform exclusive characteristic enhancement operation on the expanded sample set according to scene complexity to obtain enhanced characteristics; The fuzzy logic module is configured to utilize a fuzzy logic algorithm to perform dynamic weight calculation and feature fusion on the enhanced features based on factors influencing traffic perception, so as to obtain fusion features; And the perception prediction module is configured to carry out perception prediction on the fusion characteristics by using a perception prediction model to obtain a perception prediction result. A third aspect of the invention provides a computer