System and method for generating a surround view of a vehicle
The system adjusts imager field of view based on detected curvature to address blind spots and limited view in conventional imaging systems, achieving enhanced panoramic views with improved resolution and clarity for vehicle navigation.
Patent Information
- Application Number
- GB2025005353
- Authority / Receiving Office
- GB · GB
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-04-10
- Publication Date
- 2026-11-11
AI Technical Summary
Conventional vehicle imaging systems face challenges in capturing distant objects with high resolution and clarity due to blind spots and limited field of view, especially on curved roads, which affect applications like collision avoidance and adaptive cruise control.
A system and method that adjusts the field of view of vehicle-mounted imagers based on detected curvature in the navigation path, using sensors and deep learning to focus on regions of interest, incorporating 5-axis stabilization and sensor fusion for enhanced image capture.
The system provides an improved aerial panoramic view by capturing images over a greater distance, focusing on relevant regions, and enhancing road environment information for drivers and autonomous vehicles with higher precision and depth perception.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
TECHNICAL FIELD The present disclosure in general relates to a vehicle imaging system. More particularly, it relates to a method and system for generating a surround view of a vehicle. BACKGROUND Application of bird eye view, also referred as a panoramic view, or other assistive visual technologies like 360 degree camera view in pedestrian detection systems have become integral components in modern vehicles, as it supports functionalities such as automated parking, collision avoidance and adaptive cruise control. However, capturing distant objects using standard cameras faces significant challenges, particularly in terms of resolution and clarity of captured objects in a larger and broad view (the bird eye view). Conventional methods employ advanced image generation methods to improve the resolution and clarity of the bird eye view images. Although the conventional methods may achieve high resolutions, such conventional methods are constrained by blind spots as imaging systems in the vehicle are fixed. Thus, the conventional methods limit the correct field of view and range which is very crucial for ideal bird eye view. The blind spots become particularly challenging when generating the bird eye view in some special conditions for example, on curved roads. Furthermore, in applications like collision avoidance and adaptive cruise control, traditional systems fail due to arrangement of cameras. Thus, there is a need for an improved solution that is capable of solving the aforementioned problems of conventional imaging systems. SUMMARY Although installing multiple cameras at various positions of a vehicle facilitates generation of a surround view of the vehicle, however, it may not adequately capture region of interest relevant to a vehicle’s navigation route. Additionally, even if multiple cameras are used to capture the surround view, existing arrangement may fail to capture an extended area while operating the vehicle. Therefore, there is a need for an improved system for adapting a camera based on the vehicle’s navigation to capture the extend area while operating the vehicle. It is therefore an object of the present disclosure to provide a system and method for generating a surround view of the vehicle, for example, generating an aerial panoramic view of the vehicle also alternatively referred as bird eye view. This and other objects are achieved by means of a system, and a method defined in the appended claims. The term exemplary is in the present context to be understood as serving as an instance, example or illustration. According to an aspect of the present disclosure, a method for generating a surround view of a vehicle is disclosed. The method comprises capturing, by an imager, at least one raw frame image of a plurality of raw frame images of the vehicle’s surroundings to obtain least one two-Dimensional image (2D image) of the vehicle’s surroundings. Further, detecting a curvature in a navigation path of the vehicle based on the 2D image. The curvature is classified into at least one curvature type from a plurality of curvature types. The method further comprises adjusting a field of view of the imager in response to the detection of the curvature classified into the at least one curvature type and reconstructing an image of the curvature based on the adjustment of the field of view of the imager for obtaining the surround view of the vehicle. Optionally, performing an image plane adjustment in the imager to focus on a region of interest in the detected curvature in the at least one raw frame image. The image plane adjustment is performed by changing at least one of an orientation, position or zoomed view of the image of the curvature to be captured in the region of interest. Optionally, the plurality of curve type comprises a right curve, a left curve, an uphill or a downhill. Optionally, the classifying the at least one curvature type comprises of collecting, at least one of directional information from steering ratio of the vehicle, navigation data from Global Positioning System (GPS), and one or more positional sensor comprises of accelerometers; and calculating, a curvature rate, angular measurements, bend measurements, positional data, image processing and velocity data from the collected directional information to classify the at least one curvature type. Optionally, the classifying is performed using a deep learning method comprises of a Convolution Neural Network (CNN). Optionally, the adjusting the field of view of the imager comprises at least one of adjusting YAW in Right Direction X translation in Right Direction when the curvature type is the right curve, adjusting YAW in left Direction X translation in left Direction when the curvature type is the left curve, adjusting a pitch as Pitch Up and Y translation Upwards when the curvature type is the uphill, and adjusting the pitch as Pitch down and Y translation downwards when the curvature type is the downhill. Optionally, a magnitude of adjusting the field of view of the imager is a function of a degree of curvature. The magnitude is directly proportional to a degree of curvature. Optionally, receiving data from one or more sensors capturing information on the surround view. The image of the curvature is reconstructed by fusing the data received from the one or more sensors with the captured image of the curvature based on the adjustment of the imager. According to another aspect of the present disclosure, a system for generating a surround view of a vehicle is disclosed. The system comprises of capture, by an imager, at least one raw frame image of a plurality of raw frame images of the vehicle’s surroundings to obtain least one two-Dimensional (2D) image of the vehicle’s surroundings. The system further comprises of a processing circuitry configured to detect a curvature in a navigation path of the vehicle based on the 2D image. The curvature is classified into at least one curvature type from a plurality of curvature types. Further, the processing circuitry is configured to adjust, through an actuator, a field of view of the imager in response to the detection of the curvature classified into the curvature type; and reconstruct, an image of the curvature based on the adjustment of the field of view of the imager to obtain the surround view of the vehicle. Optionally, the processing circuitry is further configured to perform, through a sensor, an image plane adjustment in the imager to focus on a region of interest in the detected curvature. The sensor is mounted on the actuator, wherein the image plane adjustment is performed by changing at least one an orientation, position or zoomed view of the image of the curvature to be captured. Optionally, the processing circuitry is configured to adjust through the actuator, the field of view of the imager by at least one of: adjust, YAW in Right Direction X translation in Right Direction when the curvature type is the right curve, adjust, YAW in left Direction X translation in left Direction when the curvature type is the left curve, adjust, a pitch as Pitch Up and Y translation Upwards when the curvature type is the uphill and adjust, the pitch as Pitch down and Y translation downwards when the curvature type is the downhill. The actuator comprises of Micro-Electro-Mechanical Systems (MEMS) actuator. Optionally, a magnitude of adjusting the field of view of the imager is a function of a degree of curvature. The magnitude is directly proportional to a degree of curvature. Optionally, the processing circuitry is arranged to receive data from one or more sensors capturing information on the surround view, wherein the image of the curvature is reconstructed by fusing the data received from one or more sensors with the captured image of the curvature based on the adjustment of the imager. According to another aspect of the present disclosure, there is provided a computer program when loaded and run on a system, causes a processing circuitry to perform corresponding steps of method for generating a surround view of a vehicle. According to another aspect of the present disclosure, there is provided a computer-readable medium having stored thereon a computer program. Some embodiments disclosed herein have one or more of the following advantages: - The proposed system and method significantly improve ariel view range of the vehicle as it captures the images over a greater distance by adjusting the position of the sensor. - The proposed system and method provide an ability to tilt the sensor to enable the camera to focus on a region of interest relevant to vehicle navigation. - The proposed system and method provide more information on the road environment enabling drivers or autonomous vehicles for anticipated route changes. - The proposed system and method provide an enhanced platform in camera industry such as 5-Axis Stabilization (Pitch, Yaw, Roll, X&Y Translation), accelerated functionality, lower power consumption and higher precision. - The proposed system and method include a built-in 2-axis accelerometer for enhancing motion detection capabilities and compatible with existing Autofocus (AF), shutter, and zoom actuators. Further, it supports 3-Dimensioanl (3D) capture and super-resolution for providing exceptional image quality and depth perception. Other advantages may be readily apparent to one having skill in the art. Certain embodiments may have none, some, or all of the recited advantages. BRIEF DESCRIPTION OF THE DRAWINGS The foregoing will be apparent from the following more particular description of the example embodiments, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the example embodiments. FIG. 1 discloses a block diagram of a system for generating a surround view of a vehicle, according to some embodiments herein; FIG. 2 discloses a flow chart for a method for generating the surround view of the vehicle, according to some embodiments herein; FIG. 3 illustrates a working example for the system and the method for generating the surround view of the vehicle, according to some embodiments herein; FIG. 4a and 4b discloses the embodiments of an imager configured to capture an image based on the generation of the surround view through the system of FIG. 1, according to some embodiments herein; FIG 5 discloses the exemplary embodiment of the imager and an actuator, according to some embodiments herein; FIG.s 6a and 6b illustrate examples of an image sensor and the actuator, according to some embodiments herein; and FIG. 7 illustrates an example-computing environment implementing the system, as shown in FIG. 1, according to some embodiments herein. DETAILED DESCRIPTION Aspects of the present disclosure will be described more fully hereinafter with reference to the accompanying drawings. The systems and methods disclosed herein can, however, be realized in many different forms and should not be construed as being limited to the aspects set forth herein. Like numbers in the drawings refer to like elements throughout. The terminology used herein is for the purpose of describing particular aspects of the disclosure only and is not intended to limit the invention. It should be emphasized that the term “comprises / comprising” when used in this specification is taken to specify the presence of stated features, integers, steps, or components, but does not preclude the presence or addition of one or more other features, integers, steps, components, or groups thereof. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. Embodiments of the present disclosure will be described and exemplified more fully hereinafter with reference to the accompanying drawings. The solutions disclosed herein can, however, be realized in many different forms and should not be construed as being limited to the embodiments set forth herein. It will be appreciated that when the present disclosure is described in terms of a system and a method, it may also be embodied in one or more processors and one or more memories coupled to the one or more processors, wherein the one or more memories store one or more programs that perform the steps, services and functions disclosed herein when executed by the one or more processors. In an embodiment, FIG.1 discloses a block diagram of a system 100 for generating a surround view of a vehicle 180 (shown later in FIG. 3). In an example, the system 100 may be configured in the vehicle 180. The system 100 comprises processing circuitry 130 with memory 140. For example, the processing circuitry 130 and the memory 140 may be configured in the vehicle 180. Further, the processing circuitry 130 is communicatively coupled to an imager 110, an actuator 120, a steering sensor 150, a positional sensor 160 and a Global Positioning System (GPS) 170. The processing circuitry 130 is configured to receive data or signals from at least one of: the imager 110, the steering sensor 150, the positional sensors 160 and the GPS 170. In an example, the positional sensor 160 comprises one or more accelerators, gyroscope and inclinometer. The one or more positional sensors 160 are strategically mounted at specific portions of the vehicle 180 and configured to transmit motion signals to the processing circuitry 130. In an example, the one or more accelerometers may be positioned in the actuator 120 or in the vehicle 180. In a further example, the imager 110 is a visual device such as camera strategically mounted and positioned around the vehicle 180 such as at the front, left side, right side and rear sides. The imager 110 is configured to capture the surroundings of the vehicle 180 and the captured images are further processed by the processing circuitry 130 to generate an aerial view of the vehicle 180. Referring again to FIG. 1, the system 100 for generating the surround view of the vehicle 180 is shown. The system 100 comprises the imager 110 to capture at least one raw frame image of a plurality of raw frame images of the vehicle’s surroundings to obtain least one two-Dimensional image (2D) of the vehicle’s surroundings. The processing circuitry 130 is configured to detect a curvature in a navigation path of the vehicle 180 based on the 2D image. The curvature is classified into at least one curvature type from a plurality of curvature types. The processing circuitry 130 is further configured to adjust, through the actuator 120, a field of view of the imager 110 in response to the detection of the curvature classified into the curvature type and reconstruct, an image of the curvature based on the adjustment of the field of view of the imager 110 to obtain the surround view of the vehicle 180. Referring to FIG. 2, a flow chart for a method 200 for generating the surround view of the vehicle 180 is shown. The method 200 at step 210 comprises capturing, by the imager 110, at least one raw frame image of the plurality of raw frame images of the vehicle’s surroundings to obtain least one 2-Dimensional image of the vehicle’s surroundings. At step 220, the method 200 comprises detecting the curvature in the navigation path of the vehicle 180 based on the 2D image. The curvature is classified into the at least one curvature type from the plurality of curvature types. At step 230, the method 200 comprises adjusting the field of view of the imager 110 in response to the detection of the curvature classified into the at least one curvature type, and at step 240, the method 200 provides reconstructing the image of the curvature based on the adjustment of the field of view of the imager 110 for obtaining the surround view of the vehicle 180. In an embodiment, the processing circuitry 130 is further configured to perform, through the sensor 111, the image plane adjustment in the imager 110 to focus on the region of interest in the detected curvature. The sensor 111 is mounted on the actuator 120 and the image plane adjustment is performed by changing at least one of an orientation, position or zoomed view of the image of the curvature to be captured. In an example, the sensor 111 is an image sensor wherein the actuator 120 is configured to dynamically adjust the sensor 111 using five-axis stabilization, such as pitch, yaw, roll, and X&Y translation. The image plane adjustment dynamically assists in precise orientation adjustments of the sensor along the x, y, and z axis. In an example, the region of interest of the detected curvature refers to the specific segment of the vehicle's 180 navigational path where road curvature is identified or detected by curvature type from the plurality of curvature types. In an embodiment, the plurality of curve type comprises a right curve, a left curve, an uphill or a downhill. In an example, if the vehicle 180 navigates a right or left curve, the curve is detected by the steering sensor 150, positional sensors 160, and GPS 170, which then generate signals based on the vehicle’s 180 orientation. The one or more positional sensor 160 comprises one or more accelerators, gyroscope and inclinometer. The system 100 and method 200 detects the type of curve as one of the right curve or the left curve based on the received signals. In an embodiment, the method 200 further comprises classifying the at least one curvature type that comprises, collecting, at least one of directional information from steering ratio 150 of the vehicle 180, navigation data from the GPS 170, and the one or more positional sensors 160. The method 200 further comprises calculating, a curvature rate, angular measurements, bend measurements, positional data, image processing and velocity data from the collected directional information to classify the at least one curvature type. In an example, the classification refers to classifying the type of curvature as one of the right curve, the left curve, the uphill or the downhill, or in combination such as right curve uphill, right curve downhill, left curve uphill, left curve downhill based on the data inputs detected by the system 100 or the method 200. In an embodiment, the classifying is performed using deep learning method that comprises of a Convolution Neural Network (CNN). The convolutional neural network (CNN) is a multi-layered feed-forward neural network, made by stacking many hidden layers on top of each other in sequence. The sequential design may allow convolutional neural networks to learn hierarchical features. The hidden layers are typically convolutional layers followed by activation layers, some of them followed by pooling layers. The CNN may be configured to identify pattens in data. The convolutional layer may include convolutional kernels, that are used to look for patterns across the input data. The convolutional kernel may return a large positive value for a portion of the input data that matches the kernel’s pattern or may return a smaller value for another portion of the input data that does not match the kernel’s pattern. The CNN may be able to extract informative features from the training data without the need for manual processing of the training data. The CNN may produce accurate results where large unstructured data is involved, such as image classification, speech recognition and natural language processing. Also, CNN is computationally efficient, as it is able to assemble patterns of increasing complexity using the relatively small kernels in each hidden layer. The examples of CNN comprise YOLO, Trajectron etc. YOLO may be used for object detection in real-time based on bounding boxes and class probabilities. The YOLO divides an input image into a grid of cells for detecting objects within boundaries of the cells. Each grid cell may predict multiple bounding boxes and their corresponding class probabilities. In an example, the YOLO may perform image preprocessing by resizing the input image to a fixed size (e.g., 448X448) and then may divide the image into the grid of cells (7X7). The object detection may be performed by predicting multiple bounding boxes (e.g., 2-5 boxes). Each grid cell may predict class probabilities for each bounding box based on threshold values such as 0.5. Each grid cell may predict an objectness score, indicating a presence or absence of the object. A confidence score for each bounding box may be calculated by multiplying the objectness score and the class probability. The YOLO may be trained on ImageNet for classifying the curvature type based on the confidence score. In an embodiment, adjusting the field of view of the imager 110 comprises at least one of adjusting YAW in Right Direction X translation in Right Direction when the curvature type is the right curve, adjusting YAW in left Direction X translation in left Direction when the curvature type is the left curve, adjusting a pitch as Pitch Up and Y translation Upwards when the curvature type is the uphill, and adjusting the pitch as Pitch down and Y translation downwards when the curvature type is the downhill. In an embodiment, a magnitude of adjusting the the field of view of the imager 110 is a function of the degree of curvature. The magnitude is directly proportional to a degree of curvature. In an example, the imager 110 mounted on the vehicle 180 is configured to adjust its field of view more extensively when the vehicle 180 is approaching a sharper road bend to its right or left side (right curve or the left curve). Conversely, when the road exhibits a gentler or slighter curve, the imager 110 adjusts its field of view to a lesser extent to the right or the left side. In an embodiment, the system 100 and the method 200 is configured to receive data from the one or more sensors 111 capturing information on the surround view. The image of the curvature is reconstructed by fusing the data received from the one or more sensors 111 with the captured image of the curvature based on the adjustment of the imager 110. In an example, upon adjusting the position of the imager 110, the system 100 and the method 200 re-captures one raw frame image of the plurality of raw frame images having focus on the region of interest. Using sensor fusion technology, the system 100 and the method 200 integrates data from Radio Detection and Ranging, ultrasonics, Light Detection and Ranging, speedometers, and Generative Adversarial Networks. Further, the system 100 and the method 200 processes acquired images by incorporating inputs from features of the present frame along with data from long-range detection sensors enabling the generation of an enhanced surround view that focuses the regions of interest. FIG. 3 illustrates a working example 300 of the system 100 and the method 200 for generating the surround view of the vehicle 180. Initially, the one or more imager 110 may be positioned in any sides of the vehicle 180 to obtain a surround view. At step 310, the system 100 and the method 200 are configured to capture the one or more raw images at the current position of the vehicle 180. At step 320, the raw images captured are processed into undistorted Y-pyramid images. Y-Pyramid images is a processing of creating a series of images at different resolutions forming a pyramid to obtain a high-resolution image. In an example, the raw image captured by the imager 110 is distorted due to the use of wide-angle lenses and these images are corrected for geometric distortions. To correct the raw images, the image is processed through Y-pyramid generation process. During the process, the raw image is down-sampled to create multiple layers at progressively lower resolutions to correct the geometric distortions. At step 330, the processed images are classified based on the curvature in the navigation path of the vehicle 180. The curvature is classified into at least one curvature type from the plurality of curvature types. For example, the curvature may be curves (left, right) or elevation (i.e., uphill or downhill) or combination of both curves and elevation. At step 340, if the system 100 and the method 200 detects the type of curvature as the right or the left curve, the processing circuitry 130 may send a feedback signal (i.e., the degree of curvature) to the actuator 120 to adjust the imager 110 to focus on the region of interest (i.e., the type of curvature). At step 350, if the system 100 and the method 200 detects the type of curvature as the uphill or downhill, the processing circuitry 130 may send the feedback signal to the actuator 120 to adjust the imager 110 to focus on the region of interest. As the vehicle 180 progresses in any direction, the region of interest is dynamic in nature, therefore creating a new region of interest. On receiving the feedback signal, the imager 110 continuously captures new raw images focusing on updated or new region of interest which shifts according to the vehicle's 180 position. In an example, As the vehicle 180 progresses, the system 100 and the method 200 adaptively adjust the imager 110 to ensure that the imager 110 is aligned with the dynamic new region of interest. At step 360, by utilising sensor fusion technology, the aerial view (surround view) of the vehicle 180 is generated using the undistorted Y-pyramid images with the region of interest. At step 370, the aerial view of the vehicle is further enhanced using generative artificial intelligence (Al). The enhanced aerial view of the vehicle may be displayed to the user through the interface or display unit (not shown). In an example, the generative Al refers to an Al technique that uses machine learning algorithms to generate new, original and realistic content for example, images, videos, music, test, three-dimensional models (3D models) and alike. The generative Al learn patterns and structures from existing data to create new and synthetic data that resemble the original data. The generative Al here understands the pattern based on the Y-pyramid images and may generate realistic aerial view for the vehicle. In an embodiment, FIG. 4a and 4b describes the embodiments 400 of the imager 110 configured in the system 100 to capture the images of the surround view. Each imager 110 comprises the image sensor 111 configured to be dynamically mounted on a stage 410. The stage 420 is configured to be mounted on the frame 420. The stage 410 is configured to be controlled in x / y / z orientation and realize complex movements, such as pitch, yaw and roll. As shown in FIG 4b, the image sensor 111 is integrated on a Micro- Electro-Mechanical Systems (MEMS) actuator 120, driven by the 4-channel MEMS actuator 430. The actuator 120 facilitates dynamic positioning of the image sensor 111 (adjusting the field of view), allowing the image sensor 111 to move and adjust for precise positioning and stabilization of the image sensor 111 according to the region of interest. In an embodiment, FIG 5 describes an exemplary embodiment 500 of the imager 110 and the actuator 120. The image sensor 111 is dynamically mounted on the actuator 120 located in the imager 110 covered by one or more lens 510. The actuator 120 is the MEMS actuator configured to operate using a comb drive mechanism. The comb drive mechanism involves two interlocking which are electrically conductive combs that are arranged to avoid contact. If a voltage is applied to the conductive combs, electrostatic forces pull the combs towards each other, facilitating movement in a precision manner. In an example, fora 4-channel MEMS driver 430, each channel of the MEMS actuator may control movement along a specific axis. The actuator 120 is configured to move or tilt in X-direction, Y direction, Z-direction (roll or rotate), Pitch direction, Yaw direction and combination of one or more direction to focus on the region interest. Based on the feedback signal (i.e., degree of curvature) given by the processing circuitry 130, the MEMS actuator 120 will dynamically adjusts the region of interest view of the image sensor 111. In an embodiment, FIG.s 6a and 6b illustrate examples of the image sensor 111 and the actuator 120. FIG. 6a represents a previous or first position of the image sensor 111 before actuation. When the processing circuitry 130 provides the feedback signal that comprises of the degree of curvature, the actuator 120 based on the feedback signal, actuates the image sensor 111 to the desired position based on adjustment of the field of view. FIG. 6b illustrates a current or second position of the image sensor 111 after the actuation and adjustment of the field of view. FIG. 7 illustrates an example computing environment 700 implementing the system 100, as shown in FIG. 1 for generating the surround view of the vehicle 180. As depicted in FIG. 7, the computing environment 700 comprises at least one data processing circuitry 130 that is equipped with a control unit 710 and an Arithmetic Logic Unit (ALU) 720, a plurality of networking devices 740 and a plurality Input output, I / O devices 750, a memory 140, a storage 730. The data processing circuitry 130 may be responsible for implementing the system 100 and the method 200 described in FIGs 1 to 2. For example, the data processing circuitry 130 in some embodiments be equivalent to the processing circuitry 130 of the system 100 as described above in reference to FIG. 2. The data processing circuitry 130 is capable of executing software instructions stored in memory 140. The data processing circuitry 130 receives commands from the control unit 710 in order to perform its processing. Further, any logical and arithmetic operations involved in the execution of the instructions are computed with the help of the ALU 720. The computer program is loadable into the data processing circuitry 130, which may, for example, be comprised in an electronic apparatus (such as the platform). When loaded into the data processing circuitry 130, the computer program may be stored in the memory 140 associated with or comprised in the data processing circuitry 130. According to some embodiments, the computer program may, when loaded into and run by the data processing circuitry 130, cause execution of method steps according to, for example, any of the methods illustrated in FIG. 2 as described herein. The overall computing environment 700 may be composed of multiple homogeneous and / or heterogeneous cores, multiple CPUs of different kinds, special media and other accelerators. Further, the plurality of data processing circuitry 130 may be located on a single chip or over multiple chips. The algorithm comprising of instructions and codes required for the implementation are stored in either the memory 140 or the storage 730 or both. At the time of execution, the instructions may be fetched from the corresponding memory 140 and / or storage 730 and executed by the data processing circuitry 130. In case of any hardware implementations various networking devices 740 or external I / O devices 750 may be connected to the computing environment to support the implementation through the networking devices 740 and the I / O devices 750. The embodiments disclosed herein can be implemented through at least one software program running on at least one hardware device and performing network management functions to control the elements. The elements shown in FIG. 5 include blocks which can be at least one of a hardware device, or a combination of hardware device and software module. Reference signs • System 100 • Imager 110 • Image sensor 111 • Actuator 120 • Processing circuitry 130 • Memory 140 • Steering sensor 150 • Positioning sensor 160 • Global Positioning System (GPS) 170 • Vehicle 180 • Method 200 • Stage 410 • Frame 420 • 4 channel MEMS driver 430 • Lens 510 • Computing environment 700 • Control unit 710 • Arithmetic Logic Unit 720 • Storage 730 • Networking devices 740 • Input / output devices 750
Claims
1. A method (200) for generating a surround view of a vehicle (180), characterise d i n t h a t, the method 200 comprising:capturing (202), by an imager (110), at least one raw frame image of a plurality of raw frame images of the vehicle’s (180) surroundings to obtain least one two-dimensional (2D) image of the vehicle’s (180) surroundings;detecting (204) a curvature in a navigation path of the vehicle (180) based on the 2D image, wherein the curvature is classified into at least one curvature type from a plurality of curvature types;adjusting (206) a field of view of the imager (110) in response to the detection of the curvature classified into the at least one curvature type; andreconstructing (208) an image of the curvature based on the adjustment of the field of view of imager (110) for obtaining the surround view of the vehicle (180).
2. The method (200) according to claim 1, wherein the adjusting comprises:performing an image plane adjustment in the imager (110) to focus on a region of interest in the detected curvature in the at least one raw frame image,wherein the image plane adjustment is performed by changing at least one of an orientation, position or zoomed view of the image of the curvature to be captured in the region of interest.
3. The method (200) according to any of the preceding claims, wherein the plurality of curve type comprises at least one of: a right curve, a left curve, an uphill or a downhill.
4. The method (200) according to any of the preceding claims, wherein the classifying the at least one curvature type comprises:collecting, at least one of: directional information from steering ratio (150) of the vehicle (180), navigation data from Global Positioning System (GPS) (170), and one or more positional sensors (160) comprising accelerometers; andcalculating, at least one of: a curvature rate, angular measurements, bend measurements, positional data, image processing and velocity data from the collected directional information to classify the at least one curvature type.
5. The method (200) according to any of the preceding claims, wherein the classifying is performed using deep learning method comprising a Convolution Neural Network (CNN).
6. The method (200) according to any of the preceding claims, wherein the adjusting the field of view of the imager (110) comprises at least one of:adjusting YAW in Right Direction X translation in Right Direction when the curvature type is the right curve;adjusting YAW in left Direction X translation in left Direction when the curvature type is the left curve;adjusting a pitch as Pitch Up and Y translation Upwards when the curvature type is the uphill; andadjusting the pitch as Pitch down and Y translation downwards when the curvature type is the downhill.
7. The method (200) according to any of the preceding claims, wherein a magnitude of adjusting the field of view of the imager (110) is a function of a degree of curvature, wherein the magnitude is directly proportional to the degree of curvature.
8. The method (200) according to any of the preceding claims, comprising:receiving data from one or more sensors (111) capturing information on the surround view,wherein the image of the curvature is reconstructed by fusing the data received from one or more sensors (111) with the captured image of the curvature based on the adjustment of the imager (110).
9. A system (100) for generating a surround view of a vehicle (180), characterise d i n t h a t, the system (100) comprising:an imager (110) configured to capture at least one raw frame image of a plurality of raw frame images of the vehicle’s (180) surroundings to obtain least one two-dimensional (2D) image of the vehicle’s (180) surroundings;processing circuitry (130) configured to:detect a curvature in a navigation path of the vehicle (180) based on the 2D image, wherein the curvature is classified into at least one curvature type from a plurality of curvature types;adjust, through an actuator (120), a field of view of the imager (110) in response to the detection of the curvature classified into the curvature type; andreconstruct, an image of the curvature based on the adjustment of the field of view of the imager (110) to obtain the surround view of the vehicle (180).
10. The system (100) according to claim 9, wherein the processing circuitry (130) is further configured to:perform, through a sensor (111), an image plane adjustment in the imager (110) to focus on a region of interest in the detected curvature, wherein the sensor (111) is mounted on the actuator (120);wherein the image plane adjustment is performed by changing at least one an orientation, position or zoomed view of the image of the curvature to be captured.
11. The system (100) according to any of the any of the claims 9 to 10, wherein processing circuitry (130) is configured to adjust through the actuator (120), the field of view of the imager (110), wherein the processing circuitry (130) is arranged to perform at least one of:adjust, YAW in Right Direction X translation in Right Direction when the curvature type is a right curve;adjust, YAW in left Direction X translation in left Direction when the curvature type is a left curve;adjust, a pitch as Pitch Up and Y translation Upwards when the curvature type is an uphill; andadjust, the pitch as Pitch down and Y translation downwards when the curvature type is a downhill,wherein the actuator (120) comprises a Micro-Electro-Mechanical Systems (MEMS) actuator.
12. The system (100) according to any of the claims 9 to 11, wherein a magnitude of adjusting the the field of view of the imager (110) is a function of a degree of curvature, wherein the magnitude is directly proportional to the degree of curvature.
13. The system (100) according to any of the preceding claims 9 to 12, wherein the processing circuitry (130) is arranged to:receive data from one or more sensors (111) capturing information on the surround view,wherein the image of the curvature is reconstructed by fusing the data received from one or more sensors (111) with the captured image of the curvature based on the adjustment of the imager (110).
14. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out a method of any one of the claims 1 to 8.
15. A computer-readable medium having stored thereon the computer program of the claim 14.
Citation Information
Patent Citations
ViewUS2024/0100947A1onEspacenetopensinnewtab
ViewUS2021/0120158A1onEspacenetopensinnewtab
ViewUS2016/0171315A1onEspacenetopensinnewtab
ViewKR101683575B1onEspacenetopensinnewtab
ViewCN109835256AonEspacenetopensinnewtab
