阅读说明：本技术 基于回波峰值单调性分析的图像识别草地方法 (Image grassland identification method based on echo peak monotonicity analysis ) 是由 刘瑜 于 2020-05-04 设计创作，主要内容包括：本发明公开一种基于回波峰值单调性分析的图像识别草地方法,包括运动平台,所述的运动平台内部设置处理器,所述的运动平台底部安装与所述的处理器连接的特性频率为v的超声波发射模块和超声波接收模块,所述的处理器内部设置草地识别算法,包括以下步骤：(1)超声波发射模块发射一组超声波,处理器间隔ΔT时间后,将超声波接收模块输出信号进行AD采样转换为采样信号f(n),采样时间间隔Δt=1/(N·v)；当采样次数达到x·N以后,结束采样过程；(2)比较出采样信号f(n)的最大值f<Sub>max</Sub>,采样序号为n<Sub>max</Sub>；(3)获取采样信号f(n<Sub>max</Sub>+i·N)：满足|f(n<Sub>max</Sub>+i·N)|>|f(n<Sub>max</Sub>+i·N+N)|时,符合单调减规律；否则,不符合单调减规律；(4)如果f(n)不符合单调减规律,则判断为草地；否则,为非草地。(The invention discloses an image grassland identification method based on echo peak monotonicity analysis, which comprises a motion platform, wherein a processor is arranged in the motion platform, an ultrasonic transmitting module and an ultrasonic receiving module which are connected with the processor and have the characteristic frequency of v are arranged at the bottom of the motion platform, and a grassland identification algorithm is arranged in the processor, and the method comprises the following steps: (1) the ultrasonic wave transmitting module transmits a group of ultrasonic waves, after the processor separates time delta T, the processor carries out AD sampling on the output signals of the ultrasonic wave receiving module and converts the output signals into sampling signals f (N), and the sampling time interval delta T =1/(N · v); when the sampling frequency reaches x · N, ending the sampling process; (2) comparing the maximum value f of the sampled signals f (n) max Sampling number n max (ii) a (3) Acquiring a sampling signal f (n) max + i · N): satisfies | f (n) max +i·N)|>|f(n max When + i.N + N) |, the monotone decreasing rule is met; otherwise, the monotone decreasing rule is not met; (4) if f (n) does not accord with the monotone decreasing rule, judging the lawn; otherwise, it is non-grassy.)

1. The grassland image identification method based on the echo peak monotonicity analysis comprises a motion platform capable of moving autonomously, wherein a processor is arranged in the motion platform, an ultrasonic transmitting module and an ultrasonic receiving module are installed at the bottom of the motion platform, the ultrasonic transmitting module is provided with an ultrasonic transmitting probe and a signal driving circuit, the ultrasonic receiving module is provided with an ultrasonic receiving probe and a signal amplifying circuit, the signal driving circuit and the signal amplifying circuit are connected with the processor, and the characteristic frequency of the ultrasonic transmitting probe and the characteristic frequency of the ultrasonic receiving probe are v, and the grassland image identification method is characterized in that: the processor is internally provided with a grassland recognition algorithm, and the grassland recognition algorithm comprises the following steps:

(1) the processor transmits a group of ultrasonic waves with the frequency v through the ultrasonic transmitting module at regular intervals of T, and converts the output signals of the ultrasonic receiving module into sampling signals f (N) after the ultrasonic signals are transmitted at intervals of delta T, wherein N =1,2,3, and the sampling time interval delta T = 1/(N.v) is the time interval between two AD samplings; when the sampling frequency reaches x · N, ending the sampling process, wherein Δ T is the time interval from signal emission to signal sampling, N is the sampling frequency in a single ultrasonic period, and x is the number of cycles of ultrasonic signal sampling;

(2) comparing the maximum value f of the sampled signals f (n)_maxSampling number n_max；

(3) The sampling signal f (n) is obtained at an interval of 1/v of an ultrasonic wave period_max+ i · N), i =1,2,3,4, a monotonicity comparison is performed: satisfies | f (n)_max+i·N)|>|f(n_maxWhen + i.N + N) |, the monotone decreasing rule is met; otherwise, the monotone decreasing rule is not met;

(4) if the sampling signal f (n) does not accord with the monotone decreasing rule, judging the grassland; otherwise, judging the lawn is not grassland.

Technical Field

The patent relates to an image grassland identification method based on echo peak monotonicity analysis, belonging to the technical field of mobile robots.

Background

The ultrasonic wave is used as a sound wave with the frequency higher than 20000 Hz, has the advantages of good directivity, strong reflection capability and easy acquisition of concentrated sound energy, and is widely applied to occasions of distance measurement, speed measurement and the like. The ultrasonic echo signal can effectively reflect the surface characteristics of an object, and the smoother the reflection interface is, the more complete the waveform is; and multiple reflections exist in the reflection interface, so that the waveform presents a superposition form of multiple signals. The technology can be applied to automatic lawn mowers for lawn identification. Leafy many stems grass has been laid on meadow surface, for ordinary ground, carries out multiple reflection to ultrasonic signal, can distinguish meadow and ordinary ground through the morphological analysis of wave form.

Disclosure of Invention

In order to solve the problems, the patent provides an image identification grassland method based on the monotonicity analysis of echo peak values, and grassland identification is carried out by taking the attenuation characteristic of grassland or ground reflection ultrasonic signals as the characteristic.

The technical scheme adopted by the patent for solving the technical problem is as follows:

the image grassland identification method based on the echo peak monotonicity analysis comprises a motion platform capable of moving autonomously, a processor is arranged in the motion platform, an ultrasonic transmitting module and an ultrasonic receiving module are installed at the bottom of the motion platform, the ultrasonic transmitting module is provided with an ultrasonic transmitting probe and a signal driving circuit, the ultrasonic receiving module is provided with an ultrasonic receiving probe and a signal amplifying circuit, the signal driving circuit and the signal amplifying circuit are connected with the processor, the characteristic frequency of the ultrasonic transmitting probe and the characteristic frequency of the ultrasonic receiving probe are v, a grassland identification algorithm is arranged in the processor, and the grassland identification algorithm comprises the following steps:

(1) the processor transmits a group of ultrasonic waves with the frequency v through the ultrasonic transmitting module at regular intervals of T, and converts the output signals of the ultrasonic receiving module into sampling signals f (N) after the ultrasonic signals are transmitted at intervals of delta T, wherein N =1,2,3, and the sampling time interval delta T = 1/(N.v) is the time interval between two AD samplings; when the sampling frequency reaches x · N, ending the sampling process, wherein Δ T is the time interval from signal emission to signal sampling, N is the sampling frequency in a single ultrasonic period, and x is the number of cycles of ultrasonic signal sampling;

(2) comparing out samplesMaximum value f of signal f (n)_maxSampling number n_max；

(3) The sampling signal f (n) is obtained at an interval of 1/v of an ultrasonic wave period_max+ i · N), i =1,2,3,4, a monotonicity comparison is performed: satisfies | f (n)_max+i·N)|>|f(n_maxWhen + i.N + N) |, the monotone decreasing rule is met; otherwise, the monotone decreasing rule is not met;

(4) if the sampling signal f (n) does not accord with the monotone decreasing rule, judging the grassland; otherwise, judging the lawn is not grassland.

The beneficial effect of this patent mainly shows: the peak value of the reflected ultrasonic wave is extracted, the reflected waveform of the sound wave is analyzed, the monotonicity characteristic of the signal is judged, the monotonicity characteristic is used as the distinguishing characteristic of the lawn and the common ground, the lawn recognition is carried out, the non-contact detection is carried out, the reliability is high, and the anti-interference capability is strong.

Drawings

FIG. 1 is a functional block diagram of an embodiment of the present patent;

fig. 2 is a schematic diagram of an ultrasonic reflection waveform.

Detailed Description

The invention is further described below with reference to the accompanying drawings:

referring to fig. 1-2, the image grassland identification method based on the echo peak monotonicity analysis comprises a moving platform which can move autonomously, wherein the moving platform can be provided with driving wheels and supporting wheels, and can realize straight movement, backward movement and rotation at any angle. The motion platform inside set up treater 1, motion platform bottom installation ultrasonic emission module and ultrasonic wave receiving module, ultrasonic emission module set up ultrasonic emission probe 3 and signal drive circuit 2, ultrasonic receiving module set up ultrasonic receiving probe 5 and signal amplification circuit 4. The processor 1 controls the ultrasonic transmitting probe 3 to transmit ultrasonic signals through the signal driving circuit 2, the ultrasonic signals are converted into electric signals by the ultrasonic receiving probe 5 after being reflected, and the electric signals are transmitted to the processor 1 through the signal amplifying circuit 4 to be subjected to AD sampling. The characteristic frequency of the ultrasonic transmitting probe 3 and the characteristic frequency of the ultrasonic receiving probe 5 are set to be the same frequency, namely v.

The processor 1 is internally provided with a grassland identification algorithm for judging whether the current position is a grassland or not, and the grassland identification algorithm comprises the following steps:

(1) the processor 1 transmits a group of ultrasonic waves with frequency v through the ultrasonic transmitting module at regular intervals of time T, and the processor 1 converts the output signal of the ultrasonic receiving module into a sampling signal f (N) after the ultrasonic signal is transmitted at an interval of Δ T, wherein N =1,2,3, and the sampling time interval Δ T =1/(N · v), which is the time interval between two times of AD sampling; when the sampling frequency reaches x · N, ending the sampling process, wherein Δ T is the time interval from signal emission to signal sampling, N is the sampling frequency in a single ultrasonic period, and x is the number of cycles of ultrasonic signal sampling;

the processor 1 samples signals after a time interval of Δ T, so as to avoid direct interference of the ultrasonic transmitting probe 3 to the ultrasonic receiving probe 5, instead of reflected signals. The Δ T time may be set as a time required for the ultrasonic signal to travel from the ultrasonic transmission probe 3 to the ultrasonic reception probe 5. The sampling time interval delta t is set to be 1/N of the period of the ultrasonic signal, so that the sampling process and the ultrasonic signal are synchronized easily; meanwhile, x · N times the sampling time interval Δ T should be smaller than the fixed time T, i.e., the sampling should be completed within the fixed time T.

(2) Comparing the maximum value f of the sampled signals f (n)_maxSampling number n_max；

Maximum value f_maxThe peak value of the first period of the oscillation caused by the ultrasonic wave reflected from the ground at the ultrasonic wave receiving probe 5 is corresponded.

(3) The sampling signal f (n) is obtained at an interval of 1/v of an ultrasonic wave period_max+ i · N), i =1,2,3,4, a monotonicity comparison is performed: satisfies | f (n)_max+i·N)|>|f(n_maxWhen + i.N + N) |, the monotone decreasing rule is met; otherwise, the monotone decreasing rule is not met;

at the maximum value f_maxAnd sampling number n_maxThen, a peak value is acquired backward every ultrasonic period, i.e., N sampling signals. And then, comparing the amplitudes of the peaks to obtain the monotonicity characteristic of the signal. Because the grass stems, grass blades and the ground on the grassland can reflect the sound waves for multiple times, and the ultrasonic receiving probe 5 can receive the sound wave energy reflected for multiple times, the output signal can present the superposition of multiple signals, the peak value characteristic is the non-monotone decreasing characteristic, the reflected sound wave of the common ground is the primary reflection, and the signal peak value presents the monotone decreasing characteristic, so the peak value monotone characteristic can be used as the characteristic for distinguishing the grassland from the common ground.

(4) If the sampling signal f (n) does not accord with the monotone decreasing rule, judging the grassland; otherwise, judging the lawn is not grassland.

And judging whether the current position is the grassland or not by taking the peak monotonicity characteristics of the sampling signals f (n) as a judgment basis.