CN205537632U - Impact system is prevented to mobile concrete pump cantilever crane - Google Patents

Abstract

The utility model provides an impact system is prevented to mobile concrete pump cantilever crane, including wireless ultra wide band positioning unit, the three -dimensional camera unit of two meshes, controller and alarm, wireless ultra wide band positioning unit is used for utilizing wireless ultra wide band location finding technique to fix a position each articulation point of cantilever crane in real time, the three -dimensional camera unit of two meshes is used for acquireing in real time the image that contains the barrier, wireless ultra wide band positioning unit and the three -dimensional camera unit of two meshes are connected to the controller for establish space three -dimensional coordinate, process information and give -out order, alarm connection director for send alarm instruction. Compared with the prior art, the utility model provides an impact system is prevented to mobile concrete pump cantilever crane, be applicable to modern mobile concrete pump automatic pour or the intelligent control applications of other engineering machine tools in, alleviateed site work personnel's working strength to the promptness and the accuracy nature of monitoring have been improved.

Description

A kind of concrete pump truck arm collision avoidance system

Technical field

This utility model relates to collision avoidance system, a kind of anticollision system being applied to concrete pump truck arm System.

Background technology

Pump truck is as a kind of concrete conveyance and the special vehicle that pours, and its application is more and more wider.In order to meet high-rise building The construction demand in room, the jib length of concrete mixer and gradually increasing of arm joint number, length from before tens meters of development 101 meters till now, arm joint number from before 2-3 joint develop into present 4-7 joint, in casting process so that operator with There is error and time delay in the operation information transmitted between signalman by modes such as voices, is easily generated maloperation and causes collision thing Therefore.

Use infrared or laser induced carry out collision control be a kind of common technological means, generally comprise one infrared or Laser induced detector unit and a control unit, when the infrared or laser induced data detected are less than safety set in advance During value, i.e. notice control unit sends alarm or automatic control stops machine movement.This kind of device instead of manually to a certain extent Monitoring, and overcome the problems such as manual monitoring working strength is big, accuracy is low, response lag, but it is not enough to still suffer from some: (1) when pumping vehicle arm rack generation flexible bending and vibration, the error of measurement is bigger；(2) be applied to length this kind of to jib long, The excessive object of volume carries out omnibearing anticollision when controlling, and cost is high and computing is complicated；(3) be easily subject to daylight or its The interference of his close-spaced wavelength light source.Also can be by the interference of smog, dust, raindrop etc..

Thus, how to overcome above-mentioned technical problem to realize concrete pump truck arm anticollision the most efficiently and control to be one Technical problem urgently to be resolved hurrily.

Utility model content

The problems referred to above existed for prior art, the purpose of this utility model is to provide a kind of concrete the most efficiently Pumping vehicle arm rack anticollision, concrete technical scheme is as follows:

A kind of concrete pump truck arm collision avoidance system, including wireless ultra-wideband positioning unit, binocular solid image unit, Controller and alarm；Wireless ultra-wideband positioning unit is for utilizing each pass of wireless ultra-wideband location technology real-time positioning jib Node；Binocular solid image unit obtains the image containing barrier in real time；It is single that controller connects wireless ultra-wideband location Unit and binocular solid image unit, be used for setting up 3 d space coordinate, process information and sending instruction；Alarm connects control Device, is used for sending alarm command.It is preferred that 3 d space coordinate is with jib base as initial point.

Preferably, wireless ultra-wideband positioning unit includes that the location of each articulare and the jib base being installed on jib passes Sensor, alignment sensor, for transmitting location information to controller, is calculated between each articulare by controller and each joint Distance between point and base, thus draw the spatial value of each articulare of jib.

Preferably, binocular solid image unit is photographic head arranged side by side or binocular solid photographic head.

It is highly preferred that described photographic head arranged side by side or binocular solid photographic head are installed on the end of jib.Now it is easy to meter Calculate the distance between barrier and binocular solid image unit.

Preferably, controller includes graphics processing unit, computing unit and modeling unit；Described graphics processing unit is to double The barrier image information that eye stereo camera shooting unit obtains carries out image denoising process, Video Stabilization processes, image enhancement processing； Computing unit, for processing wireless ultra-wideband positioning unit and the image information of binocular solid image unit acquisition, calculates jib each Between distance and barrier and jib between the spatial value of articulare, barrier and binocular solid image unit away from The safety from whether；Modeling unit, for processing the result of calculation of computing unit and setting up the D-H space coordinates matrix of jib, builds For barrier bounding box, enclose box for the one-level bounding box of jib entirety and the secondary packet for each joint jib.By control After device processed modeling, jib defines one-level bounding box and secondary packet encloses box, and barrier also form bounding box, now computing unit Understand after receiving the image information after graphics processing unit processes and perform calculating task successively: jib one-level bounding box and barrier The distance of bounding box whether safety, as no, continue to calculate jib secondary packet and encloses the distance whether safety of box and barrier bounding box, As no, continue to calculate the minimum range whether safety of barrier and pumping vehicle arm rack.

It is highly preferred that controller also includes order unit, between barrier and jib distance less than safety away from From time send alarm command to alarm.

Compared with prior art, the concrete pump truck arm collision avoidance system that this utility model provides, it is adaptable to modern mixed In the intelligentized control method application of solidifying soil pump truck automatic pouring or Other Engineering machinery, the work alleviating field personnel is strong Degree, and improve promptness and the accuracy of monitoring.

Accompanying drawing explanation

Fig. 1 is block schematic illustration of the present utility model；

Fig. 2 is the geometric model of jib；

Fig. 3 is binocular camera shooting head model top view and three-dimensional view；

Fig. 4 is algorithm flow chart；

Fig. 5 is the geometry design sketch of barrier bounding box；

Fig. 6 is the geometry design sketch of one-level bounding box；

Fig. 7 is the geometry design sketch that secondary packet encloses box；

Fig. 8 is barrier and the geometric graph of pump truck jib distance calculating.

Detailed description of the invention

A kind of concrete pump truck arm collision avoidance system, as it is shown in figure 1, include wireless ultra-wideband positioning unit, binocular stands Body photographic head, controller and alarm.

Wireless ultra-wideband positioning unit, including being installed on each articulare and the alignment sensor of jib base of jib, Alignment sensor, for transmitting location information to controller, is calculated between each articulare by controller and each articulare and the end Distance between Zuo, thus draw the spatial value of each articulare of jib.As Fig. 2 is the geometric model of jib, at the bottom of jib Seat A₀For the initial point of solid axes, A₁～A₅Coordinate points be respectively as follows:

$\left\{\begin{array}{c}{P}_{A_1}={(x_1^t,y_1^t,z_1^t)}^T\\ P_{A_2}={(x_2^t,y_2^t,z_2^t)}^T\\ P_{A_3}={(x_3^t,y_3^t,z_3^t)}^T\\ P_{A_4}={(x_4^t,y_4^t,z_4^t)}^T\\ P_{A_5}={(x_5^t,y_5^t,z_5^t)}^T\end{array}\right.---\left(1\right)$

Wherein t represents the pumping vehicle arm rack coordinate points position at current time t.

Binocular solid photographic head is installed on the end for jib, obtains the image containing barrier in real time and is sent to control Device processed.

Controller includes graphics processing unit, computing unit, modeling unit and order unit.

Graphics processing unit is carried out at image denoising for the barrier image information obtaining binocular solid image unit Reason, Video Stabilization process, image enhancement processing.Owing to concrete mixer construction environment is complicated, obtain ring around by video camera Usually can be disturbed by various factors during the video image in border, the video image obtained can be introduced much noise so that obtain The sequence of video images obtained is unstable, reduces the accuracy of obstacle recognition.In view of the foregoing, image procossing list is therefore wanted Unit carries out the pretreatment such as image denoising, Video Stabilization to the video image that video camera obtains, it is thus achieved that without making an uproar stable image sequence Row.

Before carrying out image filtering, first the RGB Three Channel Color image collected is converted to gray-scale map, conversion meter Calculation mode is:

Gray=0.299 × R+0.587 × G+0.114 × B (2)

Being filtered gray-scale map processing, the filtering algorithm that this utility model uses is Gaussian filter algorithm, wherein Gauss The kernel function of filtering is:

$G(x,y)=\frac{1}{2{\mathrm{\π\σ}}^2}{e}^{-\frac{x^2+y^2}{2{\mathrm{\σ}}^2}}---\left(3\right)$

The process of gaussian filtering carries out convolution operation by image and gaussian kernel function and realizes, and is suppressed by gaussian filtering High fdrequency component in image, reduces image intensity value and changes greatly part faster, and remove too small details or by target Diastem couples together.

Still further aspect, photographic head is during gathering static scene, due to the random shake of photographic head so that obtain Sequence of video images unstable, original stationary object, object in the previous frame image and current frame image of video in scene Position can offset, and adds the difficulty to target recognition and analysis, is therefore being identified target in video image thing Before analyzing, video image to be carried out de-jitter, i.e. sequence of video images is carried out steady as processing.

This utility model mainly realizes sequence of video images by movement compensating algorithm based on Affine affine model Steady as processing.Assume that current acquired image relative to the motion vector of reference picture is:

a^T=(a₁,a₂,a₃,a₄,a₅,a₆) (4)

There is following transformation relation in two frames that then image sequence is adjacent:

I (x, y, t)=I (a₁x+a₂y+a₅,a₃x+a₄y+a₆,t-1) (5)

Wherein the shooting motion model of photographic head meets following mathematical formulae:

$\left[\begin{array}{c}{X}_i\\ Y_i\end{array}\right]=\left(\begin{array}{cc}{a}_1& a_2\\ a_3& a_4\end{array}\right)\left[\begin{array}{c}{X}_j\\ Y_j\end{array}\right]+\left(\begin{array}{c}{a}_5\\ a_6\end{array}\right)---\left(6\right)$

Wherein X_i,Y_iFor the coordinate of present frame impact point, X_j,Y_jCoordinate for reference frame impact point.MatrixIt is used for Represent the rotation information between two frames, scalability information and various distortion information, matrixRepresent the translation letter between two frames Breath.

According to formula 6, by calculating a₁～a₆These six parameters i.e. can determine that the affine transformation matrix of image, real Motion compensation between existing frame of video, thus reach the function of Video Stabilization.Here it is possible to by calculating between consecutive frame Little error solves a₁～a₆These six parameters, minimum error is defined as:

$E\left(a\right)=\underset{x,y\∈\mathrm{\Ω}}{\Σ}{\left[\begin{array}{c}I(x,y,t)\\ -I(a_{1}x+a_{2}y+a_5,a_{3}x+a_{4}y+a_6,t-1)\end{array}\right]}^2---\left(7\right)$

The area-of-interest of user during wherein Ω is picture frame.Formula 8 is carried out Taylor expansion, obtains equation below:

$E\left(a\right)\≈\underset{x,y\∈\mathrm{\Ω}}{\Σ}{\[s-m^{T}a\]}^2---\left(8\right)$

Wherein S=l_t+xl_x+yl_y, m^T=(xl_x,yl_x,xl_y,yl_y,l_x,l_y), l_x,l_y,l_tIt is respectively image intensity to seat Punctuate x, y and the partial derivative of time t.

According to formula 7 and formula 8, it is possible to obtain parameter a solution formula:

$a={\[\underset{\mathrm{\Ω}}{\Σ}{\mathrm{mm}}^T\]}^{-1}\[\underset{\mathrm{\Ω}}{\Σ}ms\]---\left(9\right)$

By formula 9, can be calculated so that error E approaches the vector solution of zero, thus realize the image stabilization functions of image.

Computing unit and modeling unit are the core cells that controller information processes, and both intersect and collaborative process image procossing Image information after cell processing.Wherein, computing unit is used for processing wireless ultra-wideband positioning unit and binocular solid shooting is single The information that unit obtains, calculates the spatial value of each articulare of jib, distance between barrier and binocular solid image unit And the whether safety of the distance between barrier and jib；Modeling unit is for processing the result of calculation of computing unit and setting up arm The D-H space coordinates matrix of frame, builds for barrier bounding box, for the one-level bounding box of jib entirety with for each joint The secondary packet of jib encloses box.

Fig. 3 is binocular camera shooting head model top view and three-dimensional view.According to Fig. 3, it is assumed here that in the projection of two video cameras The distance of heart line is T, and camera focus is f, and two video cameras are P (x in the same characteristic point of synchronization viewing object_c,y_c, z_c), in two video camera acquired images, the coordinate points position of target is respectively P_l(x_l,y_l) and P_r(x_r,y_r)。

Image owing to two video cameras are in sustained height, therefore impact point vertical coordinate point phase in left images With, i.e. y_l=y_r.According to prior art literature, P_l(x_l,y_l), P_r(x_r,y_r) and P (x_c,y_c,z_c) there is following relational expression:

$\left\{\begin{array}{c}{x}_l=f\·\frac{x_c}{z_c}\\ x_r=f\·\frac{x_e-T}{z_c}\\ y_l=y_r=f\·\frac{y_c}{z_c}\end{array}\right.---\left(10\right)$

The parallax of two the video camera acquired images in left and right can be represented by equation below:

D=x_l-x_r (11)

Similarity principle further according to triangle can obtain:

$d=\frac{T\×f}{D}---\left(12\right)$

According to formula 7, when the geometric position of video camera is fixed, certain point with camera light distance in the heart only by The parallax of this point determines.

Build the method that the bounding box for jib and barrier is the anticollision of a kind of practicality.Wherein sphere bounding boxes structure Make simple, and its crossing detection computation complexity is low.The determination of sphere bounding boxes has only to by arranging ball centre point coordinates (x₀,y₀,z₀) and the acquisition of radius of sphericity r.

Assume that each geometry apex coordinate of destination object is (x_i,y_i,z_i), i=1 ..., N, then the ball of sphere bounding boxes The heart is:

$\left\{\begin{array}{c}{x}_c=\frac{1}{N}\underset{i=1}{\overset{N}{\Σ}}{x}_i\\ y_c=\frac{1}{N}\underset{i=1}{\overset{N}{\Σ}}{y}_i\\ z_c=\frac{1}{N}\underset{i=1}{\overset{N}{\Σ}}{z}_i\end{array}\right.---\left(13\right)$

The distance on sphere centre coordinate and each summit is respectively d₁,d₂,...d_N, then the radius of sphere bounding boxes is:

R=max (d₁,d₂,...d_N) (14)

Test for intersection based on sphere bounding boxes is relatively simple, if the distance of two centre ofs sphere of contrast and the half of two centre ofs sphere After footpath.

The utility model proposes the multi-level bounding box anti-collision algorithm of a kind of improvement.This algorithm is a kind of novel by setting up Multi-level classification bounding box anticollision detection algorithm.Whole algorithm flow chart is illustrated in fig. 4 shown below:

First barrier is carried out the foundation of sphere bounding boxes, it is assumed that barrier is shaped as the cuboid of a standard, because of This directly selects sphere bounding boxes, its geometry design sketch as shown in Figure 5:

Due to the cuboid that barrier is rule, therefore, its sphere centre coordinate is the central point of cuboid, rectangular by calculating The meansigma methods of body apex coordinate obtains, and radius of sphericity is the cornerwise half of cuboid.

Fig. 6,7 respectively one-level bounding box and secondary packet enclose the geometry design sketch of box.Owing to pumping vehicle arm rack structure is the most multiple Miscellaneous, setting up bounding box when, it is necessary first to set up one-level bounding box, the sphere centre coordinate of one-level bounding box and the determination of radius Obtain according to formula 13 and formula 14.Secondary packet encloses the foundation of box, is primarily directed to each sub-joint of pumping vehicle arm rack and sets up bag Enclose box.Secondary packet is enclosed box and is divided according to each articulation nodes of pumping vehicle arm rack, and it is each that secondary packet encloses the sphere centre coordinate of box The meansigma methods of sub-two end points of pumping vehicle arm rack, radius is the half of this arm joint length.

This utility model proposes a kind of novel anticollision strategy according to the multi-level bounding box set up, and this strategy is concrete Step is as follows:

Step one: pumping vehicle arm rack one-level bounding box and the crossing detection of barrier bounding box, i.e. detection formula 15 are the fullest Foot condition:

$f_{xj}=\left\{\begin{array}{c}1,d_{c_1,c_2}>\left|r_1+r_2\right|\\ 0,d_{c_1,c_2}\≤\left|r_1+r_2\right|\end{array}\right.---\left(15\right)$

When two bounding box centre of sphere spacing more than two radiuses of a ball and, the most non-intersect, illustrate that pumping vehicle arm rack is safe；When two The individual bounding box centre of sphere be smaller than two radiuses of a ball and, then intersect, illustrate that pumping vehicle arm rack there may be risk of collision, strategy proceed to walk Rapid two.

The step for be equivalent to anti-collision algorithm according to a preliminary estimate, quickly whether can determine pumping vehicle arm rack by this step There is potential safety hazard, if there is hidden danger, then can quickly position concrete a certain jib potential safety hazard that may be present.

Step 2: when one-level bounding box and barrier bounding box intersect when, to pump truck nearest from intersecting area Jib is set up secondary packet and is enclosed box.Pumping vehicle arm rack secondary packet encloses the crossing detection of box and barrier bounding box, when two bounding box balls In the heart away from more than two radiuses of a ball and, the most non-intersect, illustrate that pumping vehicle arm rack is safe；When two bounding box centre ofs sphere are smaller than two The radius of a ball and, then intersect, illustrate that pumping vehicle arm rack there may be risk of collision, strategy proceeds to step 3.

By this step, it may be determined that pumping vehicle arm rack and the nearest region of obstacle distance, then according to intersecting detection knot Really, it is determined whether need computed range.

Step 3: the distance of barrier and the sub-jib of pump truck calculates, here, according to the geometric graph of Fig. 8, obstacle here The distance problem of thing and pumping vehicle arm rack can be converted to the line segment minimum range problem to a plane, due to traditional calculating side Method is extremely complex, in order to improve computational efficiency, the utility model proposes the quick distance calculating method after a kind of improvement.

It is assumed here that two node coordinates of jib with the pump truck that barrier bounding box intersects are (x₁,y₁,z₁) and (x₂,y₂, z₂), four point coordinates of EE1D1D plane areFirst the feelings that in coordinate figure, certain axle is identical are considered Condition.

Work as x₁=x₂When, the minimum range in jib and one face of barrier can be equivalent between two parallel planes Distance, now, spacing is

Work as y₁=y₂Or z₁=z₂When, the minimum range in jib and one face of barrier can be equivalent to two and hang down Distance between straight plane, now, spacing

Work as x₁≠x₂, y₁≠y₂, z₁≠z₂When, calculation is complex, needs to determine mechanical arm and obstacle simultaneously Normal between thing and the coordinate points of correspondence, this amount of calculation is bigger.Here the utility model proposes a kind of improve after near Like computational methods.Owing to the minimum range of jib and one face of barrier can be equivalent to any one plane of space and known flat The minimum range in face, now needs to calculate three distances simultaneously, then selects of minimum, and computing formula is as follows:

$d_1=\left|x_1-x_1^s\right|---\left(16\right)$

$d_2=\left|x_2-x_1^s\right|---\left(17\right)$

$d_3=|(x_1+x_2)/2-x_1^s|---\left(18\right)$

Finally, distance result of calculation is as follows:

$d=\left\{\begin{array}{cc}min(d_1,d_2,d_3),& min(x_1,x_2)>x_1^s\\ 0,& min(x_1,x_2)\≤x_1^s\end{array}\right.---\left(19\right)$

When jib is more than barrier coordinate points from the minimum range coordinate points of barrier when, by calculating d1, d2, d3 Minima obtain jib and the spacing of barrier, when jib is sat less than or equal to barrier from the minimum range coordinate points of barrier The when of punctuate, then indirect assignment is 0, then jib and barrier exist the probability of collision.

By above-mentioned step, it is possible to obtain the minimum range between jib and barrier, then with the peace pre-set Full distance threshold value contrasts, and if less than secure threshold, then reports to the police, and if greater than safe distance, then operates safety, does not carries out Report to the police.

Claims (6)

1. a concrete pump truck arm collision avoidance system, it is characterised in that: include wireless ultra-wideband positioning unit, binocular solid Image unit, controller and alarm；Wireless ultra-wideband positioning unit is used for utilizing wireless ultra-wideband location technology real-time positioning Each articulare of jib；Binocular solid image unit obtains the image containing barrier in real time；Controller connects wireless super Broadband positioning unit and binocular solid image unit, be used for setting up 3 d space coordinate, process information and sending instruction；Alarm Connect controller, be used for sending alarm command.

System the most according to claim 1, it is characterised in that: wireless ultra-wideband positioning unit includes being installed on each of jib Articulare and the alignment sensor of jib base, alignment sensor is for transmitting location information to controller, by controller Calculate the distance between each articulare and between each articulare and base, thus draw the space coordinates of each articulare of jib Value.

System the most according to claim 1, it is characterised in that: described binocular solid image unit be photographic head arranged side by side or Binocular solid photographic head.

System the most according to claim 3, it is characterised in that: described photographic head arranged side by side or binocular solid photographic head are installed End in jib.

System the most according to claim 1, it is characterised in that: controller includes graphics processing unit, computing unit and builds Form unit；The barrier image information that binocular solid image unit is obtained by described graphics processing unit is carried out at image denoising Reason, Video Stabilization process, image enhancement processing；Computing unit is used for processing wireless ultra-wideband positioning unit and binocular solid shooting The image information that unit obtains, calculates the spatial value of each articulare of jib, between barrier and binocular solid image unit Distance and barrier and jib between distance whether safety；Modeling unit is for processing the result of calculation of computing unit also Set up the D-H space coordinates matrix of jib, build for barrier bounding box, for the overall one-level bounding box of jib and for The secondary packet of each joint jib encloses box.

System the most according to claim 5, it is characterised in that: controller also includes order unit, for when barrier with Alarm command is sent to alarm when distance between jib is less than safe distance.