Multi-source signal navigation system and method for agricultural land vehicles

Abstract

The present invention discloses a multi-source signal navigation system and method for agricultural land vehicles. The system includes: a GNSS signal source receiving component, a point cloud signal source generating component, a vehicle navigation component, a point cloud density monitoring component, a positioning monitoring component, and a navigation signal source switching component. The navigation signal source switching component switches the vehicle's navigation signal source from point cloud information to GNSS information or switches the vehicle's navigation signal source from GNSS information to point cloud information based on the point cloud density monitored in real time by the point cloud density monitoring component and the positioning monitoring result of the vehicle by the positioning monitoring component, so that the vehicle navigation component can obtain the real-time positioning of the vehicle from the GNSS receiving component or the point cloud signal source generating component, thereby guiding the vehicle's traveling route.

Description

Technical Field

[0001] The present disclosure relates to a vehicle navigation method and system. More specifically, the present disclosure relates to a method and system for positioning and navigating a vehicle within a ridge of agricultural land. Background Art

[0002] With the large-scale development of agricultural planting, people's demands for agricultural mechanization and intelligence are increasing. Therefore, people need to realize the intelligence of agricultural machinery in large-scale agricultural planting, which makes it a practical need to achieve autonomous driving of vehicles in large-scale agricultural land.

[0003] Using satellite positioning (GNSS), such as Beidou navigation and GPS navigation, for vehicle autonomous driving positioning and navigation in agricultural land is a commonly used positioning and navigation measure at present. However, in agricultural land, especially in agricultural land with tall fruit trees, due to the presence of high-density leaves and branches, the positioning signal cannot be received or is weak in the ridge environment of agricultural land where occlusion or multipath refraction effects occur, and positioning interruptions often occur, which makes it difficult to stably achieve autonomous driving based on GNSS signals in agricultural land.

[0004] On the other hand, due to seasonal changes, differences in the growth status of agricultural crops in local positions, and differences between different agricultural crops, using a single navigation method will result in navigation positioning deviation and poor generality of the navigation system, thus bringing extremely poor navigation effects.

[0005] In addition, in common positioning and navigation technologies, people usually identify specific features in images for positioning, such as identifying road boundary lines in images. Due to the structured characteristics of highway roads and the existence of relatively clear and stable road signs, this feature recognition method is usually applied to the positioning of passenger car autonomous driving. Therefore, some people also try to use this feature recognition method to eliminate the instability problem of positioning and navigation using GNSS signals in tall agricultural land. However, agricultural land is a semi-structured working environment, and image features will change with light changes, seasonal changes, and also with plant growth and species changes. Therefore, the method of relying on image features for environmental recognition for positioning often cannot work stably across time and space. Once light changes, seasons change, and plant growth or fruit tree species change, a new environment will be formed, so a large amount of manual annotation and re-training of algorithms are required to ensure the accuracy of positioning.

[0006] For this reason, people also adopt the method of generating a navigation map for the entire agricultural land at one time and performing positioning and navigation based on the generated navigation map, similar to a floor cleaning robot that forms a map of the entire space to be cleaned at one time. However, due to the vastness of agricultural land and the different physical spaces formed by fruit trees as they grow and change with seasons, users need to continuously update the map. On the one hand, this requires the vehicle to have a huge map storage space and frequent map updates. On the other hand, this frequent map update process places a great operational burden on users.

[0007] Therefore, for the owners of large-scale plantations with tall plants, there is an urgent need for a simple, easy-to-implement, and low-cost method and system for a vehicle to select the best navigation source for positioning and navigation when driving automatically in the plantation. Summary of the Invention

[0008] For this reason, to solve the above technical problems, the inventors of the present disclosure recognize the respective advantages of point cloud image positioning and navigation and GNSS source navigation. Therefore, a navigation system and method capable of switching between the two in real time to obtain the best navigation source are provided. For this purpose, the present disclosure provides a multi-source signal navigation system for agricultural land vehicles, including: a GNSS source receiving component for receiving GNSS information in real time; a point cloud source generating component for obtaining point cloud information within a predetermined range in front of the vehicle in real time; a vehicle navigation component for positioning the vehicle based on GNSS information or positioning the agricultural land ridges and the vehicle based on point cloud information, so as to guide the vehicle's traveling route; a point cloud density monitoring component for monitoring the density of the obtained point cloud in real time and determining whether the density of the obtained point cloud is less than a predetermined point cloud density threshold within a predetermined time period or a continuous predetermined number of image frames; a positioning monitoring component for determining whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the agricultural land ridges displayed by the agricultural land ridge positioning is within a predetermined width range when the point cloud density monitoring component determines that the point cloud density is greater than the predetermined point cloud density threshold; and a navigation source switching component for switching the vehicle's navigation source from point cloud information to GNSS information when the point cloud density is less than the predetermined point cloud density threshold, the vehicle positioning angle deviates from the agricultural land ridge direction by more than the predetermined angle threshold, or the width of the agricultural land ridges exceeds the predetermined ridge width range, so that the vehicle navigation component can obtain the real-time positioning of the vehicle from the GNSS receiving component, thereby guiding the vehicle's traveling route.

[0009] According to the multi-source signal navigation system for agricultural land vehicles of the present disclosure, the point cloud density monitoring component includes: a first initial statistics unit that, at the vehicle startup stage, presents a confirmation option to the user and statistically calculates the initial point cloud quantity P in each frame of the point cloud image, as well as the initial left and right point cloud quantities l and r in each frame of the point cloud image, based on a predetermined number of frames of point cloud images within a predetermined time period before or after the user selects the confirmation option, so as to obtain the left and right two point cloud quantities P of the predetermined number of framesL and P R And the initial average height h of the left and right point sets of each frame of the point cloud image l and h r ; The first benchmark calculation unit calculates the average of the initial left and right point cloud numbers l and r based on the number of left and right point sets of the point cloud images of a predetermined number of frames and And the standard deviation of the number of left and right point clouds σ l and σ r , and the initial average height h of the left and right point sets of the point cloud images based on the predetermined number of frames l and h r Calculate the average height of the left and right point sets and And the standard deviation of left and right height and and a point cloud information real-time judgment unit, which is used to obtain the real-time point cloud image frame in front of the vehicle after the initialization stage, and count the number of real-time left and right point clouds in each point cloud image frame. t With r t And the real-time average height of the left and right point sets in each point cloud image frame and And in real time, the number of point clouds l t Than the average number of left point cloud The standard deviation of the left point cloud number σ below the first predetermined number l , the number of right point clouds r t The average number of right point clouds The standard deviation of the number of right point clouds σ is lower than the second predetermined number r , it is determined that the point cloud density of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud signal source generation component is not enough to provide stable positioning for the vehicle, or the real-time average height of the left point set is The average height of the left point set Lower third predetermined number of standard deviations of height left The real-time average height of the right point set The average height of the right point set Right height standard deviation lower than the fourth predetermined number When the vehicle navigation component determines that the point cloud height of the real-time point cloud image frame generated by the point cloud information source generation component is insufficient to provide stable positioning for the vehicle.

[0010] According to the multi-source signal navigation system for agricultural land vehicles of the present disclosure, the positioning and monitoring component includes: a second initial statistics unit, which presents confirmation options to the user during the vehicle startup phase, and statistically calculates the initial ridge width w and the agricultural land ridge direction in each frame of the point cloud image based on a predetermined number of frames of point cloud images within a predetermined time period before or after the user selects the confirmation option; a second reference calculation unit, which calculates the average ridge width and the ridge width standard deviation σ w of the point cloud images based on the initial ridge width w of the predetermined number of frames of point cloud images; and a ridge information real-time judgment unit, which is used to obtain the real-time ridge width w t and the vehicle traveling direction in the real-time point cloud image frame in front of the vehicle after the initialization phase in real time, and when the difference between the real-time ridge width w t and the average ridge width is greater than a fifth predetermined number of ridge width standard deviations σ w or the included angle between the vehicle traveling direction and the agricultural land ridge direction is greater than a predetermined angle threshold, it is determined that the point cloud information of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud signal source generation component is insufficient to provide stable positioning for the vehicle.

[0011] According to the multi-source signal navigation system for agricultural land vehicles of the present disclosure, the point cloud signal source generation component includes: an on-vehicle sensor, which is used to continuously collect point cloud image frames within a predetermined range in front of the vehicle along the center line of the agricultural land ridge where the vehicle is located within a predetermined driving distance of the vehicle; a spatial structure framework construction device, including: an image frame alignment unit, which aligns each subsequent frame of the point cloud image frame with the initial point cloud image frame through the ICP method to obtain the environmental point cloud information within the agricultural land ridge within the predetermined range; a voxelization processing unit, which performs voxelization processing on the space within the predetermined range to obtain the spatial structure framework corresponding to the predetermined range; and a point cloud frequency statistics unit, which statistically calculates the environmental point cloud frequency within the agricultural land ridge in each voxel unit of the spatial structure framework to obtain the probability of the existence of a specific target object in each voxel unit; a simulation correction unit, which simulates the offset amount in the horizontal direction perpendicular to the extension direction of the agricultural land ridge and the yaw angle combination around the extension direction of the agricultural land ridge of the real-time point cloud image frame obtained by the on-vehicle sensor in real time through the Monte Carlo method to obtain a plurality of corresponding simulated corrected real-time point cloud image frames; a selection unit, which selects the simulated corrected real-time point cloud image frame closest to the center of the agricultural land ridge of the spatial structure framework among the plurality of corresponding simulated corrected real-time point cloud image frames; and a navigation deviation correction instruction unit, which performs reverse positioning deviation correction on the vehicle by using the offset amount and yaw angle in the horizontal direction corresponding to the closest simulated corrected real-time point cloud image frame during simulation, so as to return the vehicle to the center position of the agricultural land ridge.

[0012] According to the multi-source signal navigation system for agricultural land vehicles of the present disclosure, the spatial structure framework construction device further includes: a filtering unit that performs a pass-through filtering process on the point cloud image frame to eliminate the point cloud beyond the predetermined range, as well as the sky noise point cloud and the ground noise point cloud within the predetermined range.

[0013] According to the multi-source signal navigation system for agricultural land vehicles of the present disclosure, the spatial structure framework construction device further includes: an update instruction unit that, when the time interval between the GNSS information received by the GNSS signal source receiving component entering the same agricultural land twice differs by a predetermined time interval, instructs to restart the construction of the spatial structure framework.

[0014] According to another aspect of the present disclosure, there is provided a multi-source signal navigation method for agricultural land vehicles, including: receiving GNSS information in real time through a GNSS signal source receiving component; obtaining point cloud information within a predetermined range in front of the vehicle in real time through a point cloud signal source generation component;

[0015] performing vehicle positioning by a vehicle navigation component based on the GNSS information or performing agricultural land ridge positioning and vehicle positioning based on the point cloud information, thereby guiding the vehicle travel route; monitoring the point cloud density obtained in real time through a point cloud density monitoring component, and determining whether the point cloud density obtained within a predetermined time period or a continuous predetermined number of image frames is less than a predetermined point cloud density threshold; determining whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the agricultural land ridge displayed by the agricultural land ridge positioning is within a predetermined width range through a positioning monitoring component when the point cloud density monitoring component determines that the point cloud density is greater than the predetermined point cloud density threshold; and switching the navigation signal source of the vehicle from the point cloud information to the GNSS information when the point cloud density is less than the predetermined point cloud density threshold, the vehicle positioning angle deviates from the agricultural land ridge direction by more than the predetermined angle threshold, or the agricultural land ridge width exceeds the predetermined ridge width range, so that the vehicle navigation component obtains the real-time positioning of the vehicle from the GNSS receiving component, thereby guiding the vehicle travel route.

[0016] According to the multi-source signal navigation method for agricultural land vehicles of the present disclosure, the step of monitoring the point cloud density obtained in real time through a point cloud density monitoring component and determining whether the point cloud density obtained within a predetermined time period or a continuous predetermined number of image frames is less than a predetermined point cloud density threshold includes: presenting a confirmation option to the user by a first initial statistics unit at the vehicle startup stage, and statistically calculating the initial point cloud number P in each frame of the point cloud image, the initial left and right point cloud numbers l and r in each frame of the point cloud image based on the point cloud images of a predetermined number of frames within a predetermined time period before or after the user selects the confirmation option, so as to obtain the left and right two point cloud numbers P L and P R and the initial average height h of the left and right point sets in each frame of the point cloud image l and hr ; calculating the average value of the initial left and right point cloud quantities l and r based on the number of left and right point sets in the point cloud images of a predetermined number of frames by the first reference calculation unit and the standard deviation σ of the left and right point cloud quantities l and σ r , and the initial average height h of the left and right point sets in the point cloud images of a predetermined number of frames l and h r calculating the average height of the left and right point sets and and the left and right height standard deviations and and the vehicle front real-time point cloud image frame after the initialization phase is obtained in real time by the point cloud information real-time judgment unit, and the real-time left and right point cloud quantities l t and r t in each frame of the point cloud image frame and the real-time average height of the left and right point sets in each frame of the point cloud image frame and and when the real-time left and right point cloud quantity l t is lower than the left point cloud number average value by the left point cloud number standard deviation σ of the first predetermined quantity l , the right point cloud quantity r t is lower than the right point cloud number average value by the right point cloud number standard deviation σ of the second predetermined quantity r , it is determined that the point cloud density of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle, or when the real-time average height of the left point set is lower than the average height of the left point set by the left height standard deviation of the third predetermined quantity or when the real-time average height of the right point set is lower than the average height of the right point set by the right height standard deviation of the fourth predetermined quantity , it is determined that the point cloud height of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle.

[0017] According to the multi-source signal navigation method for agricultural land vehicles of the present disclosure, when the positioning and monitoring component determines that the point cloud density is greater than a predetermined point cloud density threshold through the point cloud density monitoring component, determining whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the agricultural land ridge displayed by the agricultural land ridge positioning is within a predetermined width range includes: at the vehicle startup stage, the second initial statistics unit presents a confirmation option to the user and statistically calculates the initial ridge width w and the agricultural land ridge direction in each frame of the point cloud image based on a predetermined number of frames of the point cloud image within a predetermined time period before or after the user selects the confirmation option; the second reference calculation unit calculates the average ridge width based on the initial ridge width w of the predetermined number of frames of the point cloud image and the standard deviation of the ridge width σ w ; and the ridge information real-time judgment unit obtains the real-time ridge width w in the real-time point cloud image frame in front of the vehicle after the initialization stage in real time t and the vehicle traveling direction, and when the difference between the real-time ridge width w t and the average ridge width is greater than the fifth predetermined number of standard deviations of the ridge width σ w or the included angle between the vehicle traveling direction and the agricultural land ridge direction is greater than a predetermined angle threshold, it is determined that the point cloud information of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud signal source generation component is insufficient to provide stable positioning for the vehicle.

[0018] According to the multi-source signal navigation method for agricultural land vehicles of the present disclosure, the real-time acquisition of point cloud information within a predetermined range in front of the vehicle by the point cloud source generation component includes: continuously collecting point cloud image frames within the predetermined range of the agricultural land ridge in front of the vehicle along the center line of the agricultural land ridge where the vehicle is located by an in-vehicle sensor within a predetermined driving distance of the vehicle; constructing a spatial structure framework by a spatial structure framework construction device, including: aligning each subsequent point cloud image frame with the initial point cloud image frame by an image frame alignment unit using the ICP method to obtain the environmental point cloud information within the agricultural land ridge within the predetermined range, obtaining the spatial structure framework corresponding to the predetermined range through a voxelization processing unit by performing voxelization processing on the space within the predetermined range, and counting the frequency of environmental point clouds within the agricultural land ridge in each voxel unit of the spatial structure framework by a point cloud frequency statistics unit to obtain the probability of the existence of a specific target object in each voxel unit; obtaining a plurality of corresponding simulated corrected real-time point cloud image frames through a simulation correction unit by simulating the offset amount in the horizontal direction perpendicular to the extension direction of the agricultural land ridge and the yaw angle combination around the extension direction of the agricultural land ridge of the real-time point cloud image frame real-time acquired by the in-vehicle sensor in a Monte Carlo manner; selecting, by a selection unit, the simulated corrected real-time point cloud image frame closest to the center of the agricultural land ridge of the spatial structure framework from the plurality of corresponding simulated corrected real-time point cloud image frames; and performing reverse positioning and correction on the vehicle by a navigation deviation correction instruction unit using the offset amount and yaw angle in the horizontal direction corresponding to the closest simulated corrected real-time point cloud image frame during simulation, so that the vehicle returns to the center position of the agricultural land ridge.

[0019] According to the multi-source signal navigation method for agricultural land vehicles of the present disclosure, wherein the real-time acquisition of point cloud information within a predetermined range in front of the vehicle by the point cloud source generation component further includes: performing a pass-through filtering process on the point cloud image frame by a filtering unit to eliminate the point clouds outside the predetermined range and the sky noise point clouds and ground noise point clouds within the predetermined range.

[0020] According to the multi-source signal navigation method for agricultural land vehicles of the present disclosure, wherein the real-time acquisition of point cloud information within a predetermined range in front of the vehicle by the point cloud source generation component further includes: when the time interval between the GNSS information received by the GNSS source receiving component twice when entering the same agricultural land differs by a predetermined time interval, instructing, by an update instruction unit, to restart the construction of the spatial structure framework.

[0021] With the multi-source signal navigation system and method for agricultural land vehicles according to the present disclosure, it is possible to ensure that autonomous agricultural machinery can automatically travel at an adaptive level according to the point cloud information of crops in a larger operation area. When the growth conditions of crops in the operation area change, it is still possible to ensure that the autonomous agricultural machinery can continue to run smoothly in a timely manner through automatic switching between different navigation signals, providing a reliable safety guarantee for the ultimate implementation of autonomous driving of agricultural vehicles in a wider and more complex operation area. In addition, the present disclosure utilizes the point cloud signals generated by lidar or binocular cameras to locate the vehicle relative to the center line of agricultural ridges by establishing a spatial structure framework in agricultural land or plantations with tall plants. This technology enables the vehicle to perform automatic navigation in an environment without GNSS signals or with poor GNSS signals.

[0022] Other advantages, objectives, and features of the present invention will be partially reflected by the following description and partially understood by those skilled in the art through research and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 Shown is a schematic diagram of an application scenario according to the present disclosure.

[0024] Figure 2 Shown is a schematic diagram of the principle of the multi-source signal navigation system 100 for agricultural land vehicles according to the present disclosure.

[0025] Figure 3 Shown is a schematic diagram of the principle of the point cloud signal source generation component 120 of the multi-source signal navigation system 100 for agricultural land vehicles according to the present disclosure.

[0026] Figure 4 Shown is a schematic flowchart of a method for constructing a spatial structure framework for vehicle positioning within agricultural ridges of the multi-source signal navigation system 100 for agricultural land vehicles according to the present disclosure.

[0027] Figure 5 Shown is a schematic diagram of a point cloud image frame containing point cloud signals within a certain range collected by an in-vehicle sensor.

[0028] Figure 6 Shown is a schematic diagram of a three-dimensional example of the spatial structure framework formed by the method for constructing the spatial structure framework of the multi-source signal navigation system 100 for agricultural land vehicles according to the present disclosure.

[0029] Figure 7 Shown is a schematic diagram of a planar projection example of the spatial structure framework formed by the method for constructing the spatial structure framework of the multi-source signal navigation system 100 for agricultural land vehicles according to the present disclosure.

[0030] Figure 8The following is a schematic flowchart of a method for navigating a vehicle within a farmland ridge based on a spatial structure framework by the multi-source signal navigation system 100 for farmland vehicles according to the present disclosure. Detailed implementation manners

[0031] The present invention will be further described in detail below in conjunction with embodiments and the accompanying drawings, so that those skilled in the art can implement it according to the description in the specification.

[0032] Exemplary embodiments will be described in detail herein, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0033] The terms used in the present disclosure are for the purpose of describing specific embodiments only and are not intended to limit the present disclosure. The singular forms "a", "the", and "said" used in the present disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0034] It should be understood that although the terms first, second, third, etc. may be used in the present disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present disclosure, hereinafter, one of the two possible devices may be referred to as the first image frame or the second image frame. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0035] To enable those skilled in the art to better understand the present disclosure, the present disclosure will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0036] Figure 1 The following is a schematic diagram of the application scenario according to the present disclosure. As Figure 1As shown, the agricultural vehicle travels within the ridges of a ridged plantation with tall plants. Fruit trees or other cash crops taller than the agricultural vehicle are usually planted in the tall-plant plantation. Due to the natural growth of the tree canopies often overlapping with each other, the GNSS signals received by the vehicle will be unstable, making it difficult for the vehicle to achieve stable in-ridge positioning and navigation in such plantations. Therefore, autonomous driving based on GNSS signals becomes difficult to achieve. Similarly, due to the different characteristics and forms of fruit trees or other tall plants throughout the year, users will invest a large amount of energy and cost in repeatedly constructing farmland maps. Therefore, using visual features for positioning and navigation also becomes uneconomical and unrealistic. However, as the vehicle is used in different farmlands, or as the fruit trees or other tall plants in the farmland have different forms in different seasons, or there may be differences in the growth rates of fruit trees or plants in different parts of the same farmland, resulting in uneven heights. If point cloud information is always used for navigation, it will lead to unstable point cloud signals, resulting in extremely poor navigation positioning. At this time, using GNSS navigation will have a better effect. Therefore, in order to solve the situation where the navigation effect of point cloud information is worse than that of GNSS signal sources due to the differences between various farmlands and the morphological differences of plants in different parts within the farmland, the present disclosure first provides a navigation system and a navigation method that automatically switch navigation signal sources in order to utilize the advantageous navigation signal sources. Figure 2 The following shows a schematic diagram of the principle of the multi-source signal navigation system 100 for agricultural land vehicles according to the present disclosure. As Figure 2 shown, the multi-source signal navigation system 100 for agricultural land vehicles includes: a GNSS signal source receiving component 110, a point cloud signal source generating component 120, a vehicle navigation component 130, a point cloud density monitoring component 140, a positioning monitoring component 150, and a navigation signal source switching component 160. Generally speaking, based on the point cloud density monitored in real time by the point cloud density monitoring component 140 and the positioning monitoring result of the vehicle by the positioning monitoring component 150, the navigation signal source switching component 160 switches the navigation signal source of the vehicle from the point cloud signal source generating component 120 to the GNSS signal source receiving component 110 or switches the navigation signal source of the vehicle from the GNSS signal source receiving component 110 to the point cloud signal source generating component 120, so that the vehicle navigation component 130 can obtain the real-time positioning of the vehicle from the GNSS signal source receiving component 110 or the point cloud signal source generating component 120, thereby guiding the vehicle's travel route.

[0037] Specifically, refer to Figure 2As shown, the GNSS signal source receiving component 110 receives GNSS information from the satellite positioning system (such as the Beidou Navigation System (BDS) or the GPS system) in real time through the antenna. The point cloud signal source generating component 120 obtains the point cloud information within a predetermined range in front of the vehicle in real time. The vehicle navigation component 130 performs vehicle positioning based on the GNSS information or performs farmland ridge positioning and vehicle positioning based on the point cloud information, so as to guide the vehicle traveling route. During the process of the vehicle being navigated and traveling, the point cloud density monitoring component 140 monitors in real time the point cloud density of the point cloud information within a predetermined range in front of the vehicle obtained by the point cloud signal source generating component 120, and determines whether the point cloud density obtained within a predetermined time period or a continuous predetermined number of image frames is less than a predetermined point cloud density threshold. The predetermined time period can be the time of a frame of point cloud image, or can be, for example, 1, 2, 3 seconds or a longer time period. On the one hand, the point cloud density can be reflected by the number of point clouds at the local position in the point cloud image frame. For example, the number of point clouds on the left side or the right side of a frame of point cloud image. Because in a tall crop plantation, fruit trees and the like are planted on both sides of the ridge. Therefore, in the point cloud image, the number of point clouds on both sides is denser. Therefore, counting the number of point clouds on both the left and right sides can more accurately reflect the point cloud density. On the other hand, the degree of covering of the number of point clouds in the point cloud image on the farmland ridge can be determined from the height of the point cloud in the image frame. This also reflects the effect that the point cloud information can stably provide navigation positioning. When the point cloud height is higher, it means that the GNSS signal is more blocked, and better navigation effect can be obtained by using the point cloud signal. When it is lower than the predetermined height, the GNSS signal will be stronger.

[0038] Although good point cloud positioning can be achieved when the point cloud signal density is sufficient, in the case where the ridge width gradually becomes wider, the GNSS signal will be stronger. Therefore, the multi-source signal navigation system 100 for the vehicle of the present disclosure determines whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the farmland ridge displayed by the farmland ridge positioning is within a predetermined width range through the positioning monitoring component 150 when the point cloud density monitoring component determines that the point cloud density is greater than the predetermined point cloud density threshold. When the deviation between the vehicle traveling angle and the direction of the farmland ridge is too large or the ridge width is greater than a certain width, the GNSS signal will be stronger. For this reason, the navigation signal source switching component 160 switches the navigation signal source of the vehicle from the point cloud information to the GNSS information when the point cloud density is less than the predetermined point cloud density threshold, the vehicle positioning angle deviates from the farmland ridge direction by more than the predetermined angle threshold, or the farmland ridge width exceeds the predetermined ridge width range, so that the vehicle navigation component 130 obtains the real-time positioning of the vehicle from the GNSS signal source receiving component 110, thereby guiding the vehicle traveling route. Conversely, when the point cloud density, the vehicle traveling direction, the ridge direction, and the ridge width all meet the requirements, the navigation signal source switching component 160 will send a switching instruction to the vehicle navigation component 130 to switch the signal source from the GNSS signal source receiving component 110 to the point cloud signal source generating component 120.

[0039] See Figure 2 , the point cloud density monitoring component 140 includes a first initial statistics unit 141, a first reference calculation unit 142, and a point cloud information real-time judgment unit 143. During the vehicle startup phase, the first initial statistics unit 141 presents a confirmation option to the user and counts the initial point cloud quantity P in each frame of the point cloud image, the initial left and right point cloud quantities l and r in each frame of the point cloud image based on a predetermined number of frames of the point cloud image within a predetermined time period before or after the user selects the confirmation option, so as to obtain the left and right two point cloud quantities P L and P R and the initial average height h of the left and right point sets of each frame of the point cloud image l and h r . Usually, the user makes a preliminary confirmation based on the strength of the GNSS signal or can also make a confirmation based on the user's experience. Optionally, the point cloud density monitoring component 140 can directly perform an automatic confirmation based on historical parameters. For example, it automatically judges based on the parameters of the nearest navigation history record within a predetermined interval from the current calendar time. For example, if the current time is July 10th, and navigation was performed at the same location within 10 days before and after July 10th of the previous year, the navigation historical parameter will be used as the basis for confirming whether to enter the initial state. Because the growth conditions of the same farmland are usually similar at the same time, the judgment made is basically correct. Subsequently, the first reference calculation unit 142 calculates the average value of the initial left and right point cloud quantities l and r based on the left and right point set quantities of the point cloud image of the predetermined number of frames and and the standard deviation σ of the left and right point cloud quantities l and σ r , and based on the initial average height h of the left and right point sets of the point cloud image of the predetermined number of frames l and h r calculate the average height of the left and right point sets and and the left and right height standard deviations and Finally, the point cloud information real-time judgment unit 143 real-time obtains the real-time point cloud image frames in front of the vehicle after the initialization phase, and counts the real-time left and right point cloud quantities l t and r t in each frame of the point cloud image frame, as well as the real-time average height of the left and right point sets in each frame of the point cloud image frame and and when the real-time left and right point cloud quantity l t is lower than the average value of the left point cloud quantity by the standard deviation σ of the left point cloud quantity of the first predetermined quantity , the right point cloud quantity r l is higher than the average value of the right point cloud quantity by the standard deviation σ of the right point cloud quantity of the first predetermined quantity t , and when the real-time average height of the left point set Low standard deviation σ of the number of right point clouds in the second predetermined quantity r When it is determined that the point cloud density of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle, or when the real-time average height of the left point set is lower than the average height of the left point set by a third predetermined quantity of left height standard deviation or when the real-time average height of the right point set is lower than the average height of the right point set by a fourth predetermined quantity of right height standard deviation it is determined that the point cloud height of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle.

[0040] Generally speaking, during the driving process of the vehicle, the point cloud density monitoring component 140 continuously obtains point cloud input from the point cloud sensor (lidar or stereo vision). The vehicle assumes that there is sufficient point cloud to obtain reasonable and effective positioning information at the initial stage, which is called the initialization stage. In the initialization stage, there are n frames of point cloud data input. Each frame of point cloud P is divided into left and right point sets P L and P R , where l and r respectively represent the number of points in the left and right point sets of this frame, h l and h r respectively represent the average height of the left and right point sets of this frame. After the initialization stage is completed, the first reference calculation unit 142 calculates the average and standard deviation of the number of points in the left and right point sets of all frames σ l , σ r , and the average and standard deviation of the average height of the left and right point sets During the subsequent normal driving process, if or it can be considered that there is not enough point cloud to support obtaining effective and stable positioning information. If or it can be considered that the plant height is not enough to support obtaining effective and stable positioning information.

[0041] See Figure 2As described above, the positioning and monitoring component 150 includes a second initial statistics unit 151, a second reference calculation unit 152, and a ridge information real-time judgment unit 153. The second initial statistics unit 151 presents confirmation options to the user during the vehicle startup phase, and statistically calculates the initial ridge width w and the agricultural land ridge direction in each frame of point cloud image based on a predetermined number of frames of point cloud images within a predetermined time period before or after the user selects the confirmation option. The second reference calculation unit 152 calculates the average ridge width and the ridge width standard deviation σ w based on the initial ridge width w of the point cloud images of the predetermined number of frames. The ridge information real-time judgment unit 153 obtains the real-time ridge width w in the real-time point cloud image frame in front of the vehicle after the initialization phase t and the vehicle traveling direction in real time, and when the difference between the real-time ridge width w t and the average ridge width is greater than the fifth predetermined number of ridge width standard deviations σ w or the included angle between the vehicle traveling direction and the agricultural land ridge direction is greater than the predetermined angle threshold, it is determined that the point cloud information of the real-time point cloud image frame generated by the vehicle navigation component 130 based on the point cloud source generation component 120 is insufficient to provide stable positioning for the vehicle.

[0042] Generally speaking, during the initialization phase, the second initial statistics unit 151 continuously records the tree ridge width w after the vehicle is positioned. After the initialization phase is completed, the second reference calculation unit 152 calculates the average and standard deviation of the tree ridge width σ w . After the initialization is completed, after the point cloud density monitoring component 140 determines that the point cloud passes the point cloud density detection function, the ridge information real-time judgment unit 153 checks the effectiveness of the vehicle positioning result obtained by the positioning process and the tree ridge positioning through the positioning result rationality check function. If the detected tree ridge width does not meet or the absolute value of the positioning angle deviation of the vehicle relative to the ridge is greater than a predetermined angle (such as 20°, 25°, 30°, 35°, or 40°, etc.), it can also be considered that there is not enough point cloud to support obtaining effective and stable positioning information. Although the second initial statistics unit 151 and the first initial statistics unit 141 are separately described here, they can be the same initial statistics unit, which performs the statistical operations of both initial statistics units. Although the second reference calculation unit 152 and the first reference calculation unit 142 are separately described here, they can be the same reference calculation unit, which performs the reference calculation operations of both reference calculation units.

[0043] Although the present disclosure can adopt various existing methods for collecting point cloud information to obtain a point cloud image, in solving the problems of autonomous driving and positioning navigation in agricultural land, the inventors of the present disclosure recognize the repetitive characteristics of agricultural ridges in agricultural land and, in order to simplify the operability of positioning navigation, the present disclosure proposes a simple and user-friendly method for obtaining point cloud information, that is, by constructing a spatial structure framework to quickly and easily obtain point cloud information.

[0044] Figure 3 Shown is a schematic diagram of the principle of the point cloud information source generation component 120 of the multi-source signal navigation system 100 for agricultural vehicles according to the present disclosure. As Figure 3 shown, the spatial structure framework construction device 320 is connected to the vehicle-mounted sensor 310, the analog correction unit 330, the selection unit 340, and the navigation deviation correction instruction unit 350. The vehicle-mounted sensor 310 first continuously collects frames of point cloud images within a predetermined range in front of the vehicle along the center line of the agricultural ridge where the vehicle is located within a predetermined driving distance of the vehicle.

[0045] The spatial structure framework construction device 320 includes: an image frame alignment unit 321, a voxelization processing unit 322, and a point cloud frequency statistics unit 323. The image frame alignment unit 321 aligns each subsequent frame of the point cloud image with the initial point cloud image frame by the ICP method to obtain the environmental point cloud information within the agricultural ridge within the predetermined range. The voxelization processing unit 322 performs voxelization processing on the space within the predetermined range to obtain the spatial structure framework corresponding to the predetermined range. The point cloud frequency statistics unit 323 statistically analyzes the frequency of the environmental point cloud within the agricultural ridge in each voxel unit of the spatial structure framework to obtain the probability of the existence of a specific target object in each voxel unit. The analog correction unit 330 simulates the offset amount in the horizontal direction perpendicular to the extension direction of the agricultural ridge and the yaw angle combination around the extension direction of the agricultural ridge of the real-time point cloud image frame obtained by the vehicle-mounted sensor in real time by the Monte Carlo method to obtain a plurality of corresponding simulated corrected real-time point cloud image frames. The selection unit 340 selects the simulated corrected real-time point cloud image frame that is closest to the center of the agricultural ridge of the spatial structure framework among the plurality of corresponding simulated corrected real-time point cloud image frames. The navigation deviation correction instruction unit 340 uses the offset amount in the horizontal direction and the yaw angle corresponding to the closest simulated corrected real-time point cloud image frame during simulation to perform reverse positioning deviation correction on the vehicle, so that the vehicle returns to the center position of the agricultural ridge.

[0046] Optionally, the system for navigating a vehicle within a farmland ridge based on a spatial structure framework further includes a filtering unit (not shown). The filtering unit performs a passing-through filtering process on the point cloud image frames to eliminate the point clouds outside the predetermined range, as well as the sky noise point clouds and ground noise point clouds within the predetermined range. The system for navigating a vehicle within a farmland ridge based on a spatial structure framework according to the present disclosure further includes an update instruction unit 300. When the in-vehicle satellite positioning component detects that the time intervals between two consecutive entries of the in-vehicle sensor into the same farmland differ by a predetermined time interval, the update instruction unit 300 instructs to restart the construction of the spatial structure framework. The predetermined time interval is, for example, one week, two weeks, or one month.

[0047] Figure 4 Shown is a schematic flowchart of a method for constructing a spatial structure framework for positioning a vehicle within a farmland ridge by a spatial structure framework construction device 320 of a point cloud signal source generation component 120 of a multi-source signal navigation system 100 for a farmland vehicle according to the present disclosure. As Figure 4 shown, first, at step S121, as an agricultural vehicle initially arranged at the center of an arbitrary farmland ridge travels along the center line of the farmland ridge for a predetermined distance, an in-vehicle sensor 310 (such as an in-vehicle lidar, detailed later) continuously acquires point cloud image frames within a predetermined range in front of the vehicle. How to acquire the point cloud image frames belongs to the well-known technical means in the art and will not be elaborated here. The following Figure 5 shows a point cloud image frame containing point cloud signals within a certain range acquired by the in-vehicle sensor 310. The acquisition time interval of the point cloud image frames can be set by the user according to actual needs and can be between 0.01 seconds and 3 seconds, such as 0.1 seconds, 0.2 seconds, 0.5 seconds, 0.8 seconds, 1 second, etc. This acquisition interval time can be shorter or longer. Figure 5 Each image cloud point in the displayed point cloud image frame represents the existence of a specific physical object in the physical space corresponding to the image frame at that position in the image frame. For example, the point cloud may be the leaves, fruits, trunks of fruit trees, etc.

[0048] Since there are coordinate position differences along the extending direction of the farmland ridge between the point cloud image frames continuously acquired during the vehicle's travel, at step S122, the spatial structure framework construction device 320 described later performs coordinate alignment on the point cloud image frames continuously acquired within the predetermined distance. The coordinate alignment method is a conventional coordinate alignment method and will not be elaborated here. Although Figure 5There are only 6 point cloud image frames shown, but according to requirements, within a predetermined driving distance, the number of continuously acquired point cloud image frames can be more, or there can be only one. By aligning the coordinates of each subsequent point cloud image frame with the initial point cloud image frame each time, all continuously acquired point cloud image frames can be stacked vertically, thereby forming a three-dimensional point cloud spatial structure in front of the sensor within the predetermined driving distance. Each point cloud in this spatial structure represents the reflection signals of all physical objects within the field of view at this driving distance, indicating the existence of physical objects at the corresponding positions in this space, so as to obtain the point cloud information within the farmland ridges within the said predetermined range.

[0049] Then, at step S123, the spatial structure framework construction device 320 performs voxelization processing on the space within the said predetermined range to obtain the spatial structure framework corresponding to the said predetermined range. Through voxelization processing, the space within the predetermined range is gridified. The voxelization process itself is a commonly used technical means of spatial gridification, so it will not be elaborated here. The space after voxelization of the predetermined range is composed of some voxel units. The space corresponding to each voxel unit is a cubic space with a side length of 0.1 - 0.3 meters. If finer details are needed, the side length can be smaller, such as 0.05 meters. Usually, the side length is 0.1 meter. Although the voxelization step S123 is described in steps S121 and S122, it does not mean that the order of this step must be arranged like this. According to the construction method of the present disclosure, the voxelization process of the predetermined range can be directly specified at the initial stage of construction, and then steps S121 and S122 are carried out. Therefore, the Figure 2 description and display are not a limitation on the sequential relationship between the two. Optionally, the voxelization step S123 can be carried out while performing steps S121 and S122.

[0050] Next, at step S124, the spatial structure framework construction device 320 counts the frequency of the environmental point clouds within the farmland ridges in each voxel unit of the spatial structure framework to obtain the probability of the existence of specific target objects in each voxel unit. The frequency of point clouds in such a voxel unit represents the probability of the existence of corresponding physical objects in this voxel unit. For example, if 10 consecutive point cloud image frames are aligned and there are 8 point clouds in a voxel unit, it means that the probability of the existence of physical objects in the physical space corresponding to this voxel unit is 0.8. Thus, a spatial structure framework containing point cloud probability point cloud information is formed. Figure 6 Shown is a schematic diagram of a three-dimensional example of the spatial structure framework formed by the method for constructing the spatial structure framework of the multi-source signal navigation system 100 for farmland vehicles according to the present disclosure. And Figure 7Shown is a schematic diagram of a planar projection example of a spatial structure framework formed by a method for constructing a spatial structure framework of an agricultural land vehicle multi-source signal navigation system 100 according to the present disclosure. As Figure 6 shown, the probability of the presence of a physical object in each voxel unit can be represented by different colors or shades of color in the three-dimensional spatial structure framework. To meet the patent application requirements, different depths of gray scale are used in the present application to express different probabilities. As Figure 7 shown, in the schematic diagram of the planar spatial structure framework, the depth of gray scale of each square can be used to represent the level of the presence of a physical object in the voxel unit, or different textures can be used to represent it.

[0051] Optionally, it should be noted that during the acquisition process of the point cloud image frame, there will inevitably be some noise points that do not meet the requirements. To reduce the interference of noise in the subsequent processing process, according to the method for constructing a spatial structure framework for vehicle positioning within an agricultural land ridge of the present disclosure, after obtaining the original point cloud image frame, a pass-through filtering process can be performed on the point cloud image frame to eliminate the point cloud beyond the predetermined range and the sky noise point cloud and the ground noise point cloud within the predetermined range. The predetermined range is the range of 5-30 meters in front of the sensor.

[0052] Further, optionally, according to the method for constructing a spatial structure framework for vehicle positioning within an agricultural land ridge of the present disclosure, when the time interval between two consecutive detections by the in-vehicle satellite positioning component (not shown) that the in-vehicle sensor enters the same agricultural land differs by a predetermined time interval, the construction of the spatial structure framework can be restarted. The construction of the spatial structure framework can also be restarted when the in-vehicle satellite positioning component (not shown) detects that the in-vehicle sensor enters different agricultural lands.

[0053] Due to the repeatability of the agricultural land ridge environment, therefore, the spatial structure framework formed by short-distance acquisition can be used for the autonomous driving positioning and navigation of the entire agricultural land without establishing a navigation map of the entire agricultural land, thereby greatly reducing the storage space required for map navigation. More importantly, a spatial structure framework more suitable for the current growth state of the agricultural land can be established in real time according to the change of seasons, reducing the cost of repeatedly establishing an agricultural land map for a large area of agricultural land in different seasons each time, and reducing the operability of users.

[0054] According to another embodiment of the present disclosure, there is provided a method for Figure 8 navigating a vehicle within an agricultural land ridge based on a spatial structure framework by an agricultural land vehicle multi-source signal navigation system 100 according to the present disclosure as Figure 8 shown. Shown is a schematic flowchart of a method for navigating a vehicle within an agricultural land ridge based on a spatial structure framework by an agricultural land vehicle multi-source signal navigation system 100 according to the present disclosure. As Figure 8As shown, first, at step S40, a spatial structure framework for vehicle positioning within the agricultural land ridges is constructed using the method described above for Figure 4 Subsequently, at step S50, the in-vehicle sensor 310 is used to obtain in real time a frame of real-time point cloud image within a predetermined range in front of the agricultural land ridge where the vehicle is located. Then, at step S60, the real-time point cloud image frame is aligned with the initial point cloud image frame by the ICP method. This alignment process can be directly performed by the alignment unit in the spatial structure framework construction device 320. Optionally, the real-time point cloud image frame can also be filtered by the pass-through filtering method.

[0055] Next, at step S60, the simulation correction unit 330 obtains a plurality of corresponding simulated corrected real-time point cloud image frames by simulating, in the Monte Carlo manner, the offset amount in the horizontal direction perpendicular to the extension direction of the agricultural land ridge and the yaw angle around the extension direction of the agricultural land ridge for the real-time point cloud image frame. The Monte Carlo simulation is a commonly used simulation method, so it will not be elaborated here. Specifically, for each real-time point cloud image frame, it is assumed to be obtained when the vehicle is offset by a certain distance in the horizontal direction and deflected by a certain angle around the center line of the agricultural land ridge. Therefore, through the Monte Carlo simulation method, the real-time point cloud image frame is subjected to offset transformation and rotation transformation to form a simulated corrected real-time point cloud image frame. Since the amplitudes and combinations of the simulated offset distances or deflection angles are different, multiple simulated corrected real-time point cloud image frames will be formed for a single real-time point cloud image frame. The purpose is to try to obtain a corrected image that is similar to or the same as the point cloud image frame obtained in the correct driving direction. This provides a basis for correcting the subsequent vehicle positioning and navigation. Therefore, subsequently, at step S70, the selection unit 340 in the vehicle navigation system selects the simulated corrected real-time point cloud image frame that is closest to the center of the agricultural land ridge of the spatial structure framework among the plurality of corresponding simulated corrected real-time point cloud image frames. For example, the specific process of comparison is to sample the possible y-direction offset and yaw-direction offset through the Monte Carlo simulation to obtain N simulated sampling postures, and assume that the simulated sampling posture is the real posture to convert the currently obtained point cloud into the spatial framework structure coordinate system (y = 0, yaw = 0, if the sampling posture is the real posture, the converted point cloud should be facing the ridge center). Through comparison, the voxel units with point cloud images in the simulated corrected real-time point cloud image frames are obtained, and the corresponding voxel units in the spatial framework structure and the point cloud frequency or normalized frequency of the corresponding voxel units are found. Subsequently, the point cloud frequencies or normalized frequencies of all the found corresponding voxel units are obtained. Therefore, the point cloud probability distribution of each small grid in the gridded space of the vehicle forward direction view. The depth of each grid represents the magnitude of the probability (the probability that there is an object in the space represented by this grid is 0-1). Such as Figure 6As shown, all the point cloud voxel units combined together are actually a three-dimensional framework, such as Figure 6 shown, where the color depth of the voxel units in different parts represents different point cloud probabilities. Since the three-dimensional grid space cannot be intuitively visualized, for this reason, Figure 7 the two-dimensional view shown is used for representation. As Figure 6 shown in the three-dimensional space structure, the space is filled with points, only the probabilities are different, and the probabilities of almost all points are greater than zero. For visualization, Figure 6 voxel units with probabilities less than a certain value are not shown in. Specifically for actual objects, the probabilities of these voxel units correspond to the probabilities of all things existing in the actual space (such as the distribution of tree trunks, branches, leaves, and sensor noise points). Taking out the probability of each point of the transformed point cloud in the space framework structure, the product of all point probabilities can obtain an overall probability. Among all the sampling postures, the posture with the highest overall probability of the transformed point cloud is the best estimated posture. Optionally, in order to eliminate the influence of noise. For voxel units without point clouds in reality, a fixed probability value, such as 0.1, is assigned to them.

[0056] Finally, at step S80, the navigation deviation correction instruction unit 350 performs reverse positioning and deviation correction on the vehicle according to the horizontal offset and yaw angle corresponding to the selected closest simulated corrected real-time point cloud image frame during simulation, so that the vehicle returns to the center position of the farmland ridge. For example, if the leftward offset distance used during the simulation of the closest simulated corrected real-time point cloud image frame is 0.5 meters and the counterclockwise deflection angle is 5 degrees, then during the positioning deviation correction, the vehicle is instructed to turn clockwise by 5 degrees and walk rightward by 0.5 meters, so that the vehicle returns to the center of the farmland ridge again, realizing re-accurate positioning and correct navigation.

[0057] Optionally, according to the method for vehicle navigation in a farmland ridge based on a spatial structure framework of the present disclosure, the vehicle can also be continuously positioned through a closed-loop algorithm, and the speed and direction of the vehicle can be adjusted so that the vehicle is navigated towards the center position of the farmland ridge.

[0058] Optionally, it should be noted that when the selection unit 340 selects the closest simulated corrected real-time point cloud image frame, first, for each simulated corrected real-time point cloud image frame, the point cloud probability in each voxel unit is obtained according to the spatial structure framework; then the product of all non-zero point cloud probabilities is calculated to obtain the overall point cloud probability of each simulated corrected real-time point cloud image frame; finally, the simulated corrected real-time point cloud image frame with the highest overall point cloud probability is selected as the closest simulated corrected real-time point cloud image frame.

[0059] Furthermore, the method for vehicle navigation within farmland ridges based on the spatial structure framework of the navigation system according to the present disclosure can also measure the traveling distance of the vehicle after it enters the farmland ridge, and when the traveling distance is greater than the length of the farmland ridge, steer the vehicle to navigate to an adjacent farmland ridge.

[0060] The basic principles of the present disclosure have been described above in conjunction with specific embodiments. However, it should be noted that for those of ordinary skill in the art, all or any steps or components of the method and apparatus of the present disclosure can be implemented in any computing device (including a processor, a storage medium, etc.) or a network of computing devices in hardware, firmware, software, or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present disclosure.

[0061] Therefore, the object of the present disclosure can also be achieved by running a program or a set of programs on any computing device. The computing device can be a well-known general-purpose device. Therefore, the object of the present disclosure can also be achieved merely by providing a program product containing program code for implementing the method or apparatus. That is to say, such a program product also constitutes the present disclosure, and a storage medium storing such a program product also constitutes the present disclosure. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future.

[0062] It should also be noted that in the apparatus and method of the present disclosure, obviously, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present disclosure. And the steps of performing the above series of processes can naturally be executed in chronological order according to the described order, but it is not necessary to execute them in chronological order. Some steps can be executed in parallel or independently of each other.

[0063] The above specific embodiments do not constitute a limitation to the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. An agricultural land vehicle multi-source signal navigation system, comprising: a GNSS signal source receiving component for receiving GNSS information in real time; a point cloud signal source generating component for obtaining point cloud information within a predetermined range in front of the vehicle in real time; a vehicle navigation component for positioning the vehicle based on GNSS information or positioning the agricultural land ridges and the vehicle based on point cloud information, thereby guiding the vehicle's travel route; a point cloud density monitoring component for monitoring the density of the obtained point cloud in real time and determining whether the density of the obtained point cloud within a predetermined time period or a continuous predetermined number of image frames is less than a predetermined point cloud density threshold; a positioning monitoring component for determining whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the agricultural land ridges displayed by the agricultural land ridge positioning is within a predetermined width range when the point cloud density monitoring component determines that the point cloud density is greater than the predetermined point cloud density threshold; and a navigation signal source switching component for switching the vehicle's navigation signal source from point cloud information to GNSS information when the point cloud density is less than the predetermined point cloud density threshold, the vehicle positioning angle deviates from the agricultural land ridge direction by more than the predetermined angle threshold, or the width of the agricultural land ridges exceeds the predetermined ridge width range, so that the vehicle navigation component obtains the real-time positioning of the vehicle from the GNSS receiving component, thereby guiding the vehicle's travel route.

2. The agricultural land vehicle multi-source signal navigation system according to claim 1, wherein the point cloud density monitoring component comprises: The first initial statistical unit presents confirmation options to the user during the vehicle startup phase, and based on the point cloud images of a predetermined number of frames within a predetermined time period before or after the user selects the confirmation option, counts the initial point cloud quantity P in each frame of the point cloud image, as well as the initial left and right point cloud quantities l and r in each frame of the point cloud image, so as to obtain the left and right point cloud quantities P of the predetermined number of frames L and P R and the initial average height h of the left and right point sets of each frame of the point cloud image l and h r ; The first reference calculation unit calculates the average of the initial left and right point cloud quantities \(l\) and \(r\) based on the number of points in the left and right point sets of the point cloud images for a predetermined number of frames and the standard deviation \(\sigma\) of the left and right point cloud quantities l and \(\sigma\) r , and calculates the average height of the left and right point sets based on the initial average heights \(h_l\) and \(h_r\) of the left and right point sets of the point cloud images for a predetermined number of frames and the left and right height standard deviations and and A point cloud information real-time judgment unit is used to obtain in real time the real-time point cloud image frame in front of the vehicle after the initialization stage, and count the real-time left and right point cloud numbers l in each frame of the point cloud image frame t and r t as well as the real-time average height of the left and right point sets in each frame of the point cloud image frame and and when the real-time left and right point cloud numbers l t are lower than the standard deviation σ of the left point cloud number with the first predetermined quantity compared to the average number of left point clouds 、the right point cloud number r l is lower than the standard deviation σ of the right point cloud number with the second predetermined quantity compared to the average number of right point clouds t r it is determined that the point cloud density of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle, or when the real-time average height of the left point set is lower than the standard deviation of the left height with the third predetermined quantity compared to the average height of the left point set or when the real-time average height of the right point set is lower than the standard deviation of the right height with the fourth predetermined quantity compared to the average height of the right point set it is determined that the point cloud height of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle.​​​​ 3. The agricultural land vehicle multi-source signal navigation system according to claim 1, wherein the positioning monitoring component comprises: a second initial statistics unit, during the vehicle startup phase, presenting a confirmation option to the user and statistically calculating the initial ridge width w and the agricultural land ridge direction in each frame of point cloud image based on a predetermined number of frames of point cloud images within a predetermined time period before or after the user selects the confirmation option; The second reference calculation unit, the average ridge width calculated based on the initial ridge width w of the point cloud images of a predetermined number of frames and the ridge width standard deviation σ w ; The ridge information real-time judgment unit is used to obtain in real time the real-time ridge width w in the real-time point cloud image frame in front of the vehicle after the initialization stage t and the traveling direction of the vehicle, and when the difference between the real-time ridge width w t and the average ridge width is greater than the fifth predetermined number of ridge width standard deviations σ w or the included angle between the traveling direction of the vehicle and the direction of the farmland ridges is greater than the predetermined angle threshold, it is determined that the point cloud information of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle.

4. The agricultural land vehicle multi-source signal navigation system according to any one of claims 1-3, wherein the point cloud signal source generating component comprises: a vehicle-mounted sensor for continuously collecting frames of point cloud images within a predetermined range of the agricultural land ridge in front of the vehicle along the center line of the agricultural land ridge where the vehicle is located within a predetermined driving distance of the vehicle; a spatial structure framework construction device, comprising: an image frame alignment unit for aligning each subsequent frame of point cloud image frame with the initial point cloud image frame by the ICP method to obtain the environmental point cloud information within the agricultural land ridge within the predetermined range; a voxelization processing unit for performing voxelization processing on the space within the predetermined range to obtain the spatial structure framework corresponding to the predetermined range; and a point cloud frequency statistics unit for statistically calculating the frequency of the environmental point cloud within the agricultural land ridge in each voxel unit of the spatial structure framework to obtain the probability of the existence of a specific target object in each voxel unit; a simulation correction unit for simulating the offset amount of the real-time point cloud image frame obtained by the vehicle-mounted sensor in real time along the horizontal direction perpendicular to the extension direction of the agricultural land ridge and the yaw angle combination around the extension direction of the agricultural land ridge by the Monte Carlo method to obtain a plurality of corresponding simulated corrected real-time point cloud image frames; A selection unit that selects the simulated corrected real-time point cloud image frame among the multiple corresponding simulated corrected real-time point cloud image frames that is closest to the center of the agricultural land ridge of the spatial structure framework; and A navigation deviation correction instruction unit that uses the horizontal offset and yaw angle corresponding to the closest simulated corrected real-time point cloud image frame during simulation to perform reverse positioning and deviation correction on the vehicle, so that the vehicle returns to the center position of the agricultural land ridge.

5. The multi-source signal navigation system for agricultural vehicles according to claim 4, wherein the spatial structure framework construction device further includes: A filtering unit that performs a pass-through filtering process on the point cloud image frame to eliminate the point cloud outside the predetermined range, as well as the sky noise point cloud and ground noise point cloud within the predetermined range.

6. The multi-source signal navigation system for agricultural vehicles according to claim 4, wherein the spatial structure framework construction device further includes: An update instruction unit that, when the time interval between the GNSS information received by the GNSS signal source receiving component entering the same agricultural land twice differs by a predetermined time interval, instructs to restart the construction of the spatial structure framework.

7. A multi-source signal navigation method for agricultural vehicles, including: Receiving GNSS information in real time through a GNSS signal source receiving component; Obtaining point cloud information within a predetermined range in front of the vehicle in real time through a point cloud signal source generation component; Guiding the vehicle's travel route by performing vehicle positioning based on GNSS information or agricultural land ridge positioning and vehicle positioning based on point cloud information through a vehicle navigation component; Monitoring the point cloud density obtained in real time through a point cloud density monitoring component, and determining whether the point cloud density obtained within a predetermined time period or a continuous predetermined number of image frames is less than a predetermined point cloud density threshold; When the point cloud density monitoring component determines that the point cloud density is greater than the predetermined point cloud density threshold, determining whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the agricultural land ridge displayed by the agricultural land ridge positioning is within a predetermined width range through a positioning monitoring component; and When the point cloud density is less than the predetermined point cloud density threshold, the vehicle positioning angle deviates from the agricultural land ridge direction by more than the predetermined angle threshold, or the width of the agricultural land ridge exceeds the predetermined ridge width range, switching the vehicle's navigation signal source from point cloud information to GNSS information through a navigation signal source switching component, so that the vehicle navigation component obtains the real-time positioning of the vehicle from the GNSS receiving component, thereby guiding the vehicle's travel route.

8. The multi-source signal navigation method for agricultural vehicles according to claim 7, wherein the step of monitoring the point cloud density obtained in real time through a point cloud density monitoring component and determining whether the point cloud density obtained within a predetermined time period or a continuous predetermined number of image frames is less than a predetermined point cloud density threshold includes: In the vehicle startup phase, the first initial statistical unit presents confirmation options to the user, and based on the point cloud images of a predetermined number of frames within a predetermined time period before or after the user selects the confirmation option, it counts the initial point cloud quantity P in each frame of the point cloud image, as well as the initial left and right point cloud quantities l and r in each frame of the point cloud image, thereby obtaining the left and right point cloud quantities P of the predetermined number of frames L and P R as well as the initial average height h of the left and right point sets of each frame of the point cloud image l and h r ; The first reference calculation unit calculates the average value of the initial left and right point cloud quantities l and r based on the number of points in the left and right point sets of the point cloud images for a predetermined number of frames. and the standard deviation σ of the left and right point cloud quantities l and σ r , and the initial average height h of the left and right point sets of the point cloud images for a predetermined number of frames l and h r calculate the average height of the left and right point sets and and the left and right height standard deviations and and The point cloud information real-time judgment unit obtains the real-time point cloud image frame in front of the vehicle after the initialization stage, and counts the number of real-time left and right point clouds in each point cloud image frame. t With r t And the real-time average height of the left and right point sets in each point cloud image frame and And in real time, the number of point clouds l t Than the average number of left point cloud The standard deviation of the left point cloud number σ below the first predetermined number l , the number of right point clouds r t The average number of right point clouds The standard deviation of the number of right point clouds below the second predetermined number σ r , it is determined that the point cloud density of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud signal source generation component is not enough to provide stable positioning for the vehicle, or the real-time average height of the left point set is The average height of the left point set Lower third predetermined number of standard deviations of height left The real-time average height of the right point set The average height of the right point set Right height standard deviation lower than the fourth predetermined number When the vehicle navigation component determines that the point cloud height of the real-time point cloud image frame generated by the point cloud information source generation component is insufficient to provide stable positioning for the vehicle.

9. The multi-source signal navigation method for agricultural vehicles according to claim 7, wherein the step of determining whether the vehicle positioning angle deviation is greater than a predetermined angle threshold or whether the width of the agricultural land ridge displayed by the agricultural land ridge positioning is within a predetermined width range when the point cloud density monitoring component determines that the point cloud density is greater than the predetermined point cloud density threshold through a positioning monitoring component includes: During the vehicle startup phase, the second initial statistical unit presents confirmation options to the user, and based on the point cloud images of a predetermined number of frames within a predetermined time period before or after the user selects the confirmation option, it statistically analyzes the initial ridge width w and the agricultural land ridge direction in each frame of the point cloud image. The average ridge width calculated by the second reference calculation unit based on the initial ridge width w of the point cloud images for a predetermined number of frames and the ridge width standard deviation σ w ; and The real-time ridge width w in the real-time point cloud image frame in front of the vehicle after the initialization stage is obtained in real time through the ridge information real-time judgment unit t and the vehicle traveling direction, and when the real-time ridge width w t is greater than the standard deviation σ of the ridge width of the fifth predetermined quantity between the average ridge width or the included angle between the vehicle traveling direction and the farmland ridge direction is greater than the predetermined angle threshold, it is determined that the point cloud information of the real-time point cloud image frame generated by the vehicle navigation component based on the point cloud source generation component is insufficient to provide stable positioning for the vehicle. w ​ 10. The agricultural land vehicle multi-source signal navigation method according to any one of claims 7-9, wherein the point cloud information within a predetermined range in front of the vehicle is obtained in real time by the point cloud source generation component includes: Continuously collecting frames of point cloud images within a predetermined range of the agricultural land ridge in front of the vehicle along the center line of the agricultural land ridge where the vehicle is located by an in-vehicle sensor within a predetermined driving distance of the vehicle. Constructing a spatial structure framework by a spatial structure framework construction device, including: making subsequent frames of point cloud images align with the initial point cloud image frame by the image frame alignment unit using the ICP method to obtain the environmental point cloud information within the agricultural land ridge within the predetermined range, obtaining the spatial structure framework corresponding to the predetermined range through voxelization processing of the space within the predetermined range by the voxelization processing unit, and statistically analyzing the environmental point cloud frequency within the agricultural land ridge in each voxel unit of the spatial structure framework by the point cloud frequency statistical unit to obtain the probability of the existence of a specific target object in each voxel unit. Obtaining a plurality of corresponding simulated corrected real-time point cloud image frames through the simulation correction unit by simulating the offset amount in the horizontal direction perpendicular to the extension direction of the agricultural land ridge and the yaw angle combination around the extension direction of the agricultural land ridge of the real-time point cloud image frames obtained in real time by the in-vehicle sensor in a Monte Carlo manner. Selecting the simulated corrected real-time point cloud image frame closest to the center of the agricultural land ridge of the spatial structure framework from the plurality of corresponding simulated corrected real-time point cloud image frames by the selection unit. and Using the offset amount in the horizontal direction and the yaw angle corresponding to the closest simulated corrected real-time point cloud image frame during simulation by the navigation deviation correction instruction unit to perform reverse positioning deviation correction on the vehicle, so that the vehicle returns to the center position of the agricultural land ridge.

11. The agricultural land vehicle multi-source signal navigation method according to claim 10, wherein the obtaining of the point cloud information within a predetermined range in front of the vehicle by the point cloud source generation component in real time further includes: Performing direct-pass filtering processing on the point cloud image frames by the filtering unit to eliminate the point clouds outside the predetermined range, as well as the sky noise point clouds and ground noise point clouds within the predetermined range.

12. The agricultural land vehicle multi-source signal navigation method according to claim 10, wherein the obtaining of the point cloud information within a predetermined range in front of the vehicle by the point cloud source generation component in real time further includes: When the time interval between the GNSS information received by the GNSS source receiving component entering the same agricultural land twice differs by a predetermined time interval, the update instruction unit instructs to restart the construction of the spatial structure framework.

Images