JP2709301B2 - Striation light extraction circuit - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a system which uses position of a plurality of lines of light on a television image to obtain position information by processing an image signal in real time. The present invention relates to a light extraction circuit. [Prior Art] Conventionally, as one form of vision of a robot, a distance information input device (range finder) for projecting striation light onto an object surface and measuring the shape of the object based on the principle of triangulation has been known (Oshima). , “Study of Object Recognition Using Three-Dimensional Information”, Research Report of AIST, No. 826 (Showa 57)). In this type of apparatus, it is necessary to determine the position of the linear light projected on the object on the TV image in order to obtain the distance to the target object. In this type of apparatus, it is necessary to perform data preprocessing. The reason is that the amount of information of the luminance data of the entire image becomes very large (about 256 KB per screen)
That's why. In addition, a large storage medium and a long transfer time are required to store image information for many screens. [Problems to be Solved by the Invention] In this type of conventional device, in order to determine the position of one line of light, a point where the luminance of the image signal is maximum in each horizontal line is searched for. Using a preprocessing circuit that sets the midpoint between the two points where the slope on both sides is the largest, and the position of the striation light
Real-time processing has been done. However, in the above pre-processing circuit, it is sufficient if at most one line of light is present in each horizontal line, but a distance information input device for projecting a plurality of lines of light (Japanese Patent Laid-Open No. 62-1987).
For example, in the case where the method is applied to Japanese Patent Application Laid-Open No. 63804, the position of a plurality of striations in an image must be extracted. Accordingly, an object of the present invention is to provide a striation light extraction circuit that can determine a plurality of striation light positions on a TV image in real time in view of the above points. [Means for Solving the Problems] In order to achieve the above object, according to the present invention, there is provided a means for obtaining a protruding signal component appearing on an image signal representing a plurality of striations as a candidate for striations, Means for providing an evaluation function value for eliminating the streak light candidate for each of the streak light candidates, the previously obtained evaluation function value, and the newly obtained evaluation function value A means for comparing, and a means for leaving only a predetermined number of linear light candidates as a result of the comparison, processing the image signal at a video rate. Further, the linear light extraction circuit according to the present invention uses the luminance information V and the time information T of the pixel points constituting the candidate signal from the image signal candidate for the linear light, A weighted average value calculation circuit for determining the streak position by the calculation of (V · T) / ΣV. [Operation] In the present invention, unnecessary linear light candidates are removed using a predetermined evaluation function, and the position of the linear light is determined by performing a weighted operation. EXAMPLES Hereinafter, the present invention will be described in detail based on examples. FIG. 1 is a diagram showing the luminance level of the streak light displayed on the TV image. As shown in the figure, as a method of obtaining the position of the linear light, in this embodiment, a portion exceeding the threshold value is detected as one block, the center of gravity of the peak is obtained, and this is set as the position of the linear light. This method of determining the position of the striation light uses a plurality of data with high luminance to determine the position of one striation light, so that the position can be determined without error even when high frequency noise is mixed. It is. Now, assuming that the time axis in FIG. 1 is T and the luminance is V, the position of the center of gravity Tg is expressed as in equation (1). Tg = To + Σ (T′V) / (ΣV) (1) where To is the time when the luminance exceeds the threshold value Vo, and T ′ is the value of T−To. In this specification, capital letters of the alphabet are used.
When suffixes (subscripts at the lower right) are added to T, V, S, and K, half-width characters (that is, the width of the characters are halved) are uniformly used as these capital letters. FIG. 2 is a block diagram showing the configuration of the entire embodiment. In the figure, reference numeral 2 denotes an A / D conversion unit, which converts a luminance signal representing linear light into a digital value. Reference numeral 4 denotes a threshold value calculator for inputting digitized luminance information V for one horizontal scanning line and determining a threshold value. Reference numeral 6 denotes a one-line delay unit, and reference numeral 8 denotes line light detection and parameter calculation / selection (to be described in detail later).
An arithmetic unit 10 for performing the above operation is a center-of-gravity calculating unit that determines the position of the striation light. The circuit shown in the figure is configured to perform signal processing in three stages in order to perform high-speed processing. That is, in each stage, data for one horizontal line is processed within the time of one horizontal line, and the result is sequentially passed to the next stage. In each of these steps, the following processing is performed. (1) Calculation of threshold: The sum of luminance data of each pixel on a horizontal line is obtained, and a value obtained by dividing the sum by 256 (half of the number of pixels in the horizontal direction) is set as a threshold Vo. (2) Parameter calculation and selection: For each striae, To,
Calculate ′ (T′V) and ΣV. At this time, the magnitude of the newly obtained ΔV (used as the evaluation function value) is compared with the N ΔV values obtained in the past, and if the new value is larger than at least one of the old values, the minimum Is replaced with a new one. (3) Calculation of the center of gravity: Using the obtained N sets of parameters, N center of gravity is calculated from equation (1). When performing the calculation, in order to shorten the division time, an LUT of 1 / ΔV data is prepared in the ROM in advance, and the value is multiplied by referring to the value with ΔV. Among these, a particularly important process is the sorting performed in (2). This is necessary in order to distinguish between streaking and noise due to a plurality of peaks exceeding a threshold value on one horizontal line. In general, the area of the peak due to noise is considered to be small. Therefore, in this embodiment, the value of ΔV necessary for calculating the position of the center of gravity is used as an evaluation function value and used for the determination. FIG. 3 is a block diagram showing a part of the arithmetic unit 8 shown in FIG. 2 in detail. That is, this drawing is a block diagram for calculating parameters necessary for obtaining the center of gravity of the linear light waveform shown in FIG. Each symbol shown in the figure has the following meaning. V: luminance signal, Vo: threshold luminance, φ: clock, T: time from the start of horizontal scanning (horizontal coordinate value in FIG. 1), T ′: elapsed time from the detection of linear light (T ′ = T) -To) S: calculation result of ΣV (evaluation function value), ST ′: calculation result of Σ (VT ′), To: time when streak is detected (see FIG. 1). The circuit shown in FIG. 3 performs the following operation. a) The value of the luminance V and the threshold luminance Vo are compared in the comparator 12. When V becomes greater than Vo, comparator 12 issues a trigger signal 12A. b) With the generation of the trigger signal 12A, T '
Output from At the same time, the calculation of ΣV, Σ (VT ') is performed in each of the calculation units 16 and 18. Further, the time To at the start of the calculation is stored in the latch circuit 20. c) When the luminance V becomes smaller than the threshold luminance Vo again, the above calculation is terminated, and each calculation value is output. FIG. 4 is a block diagram showing a circuit for performing sorting (that is, for selecting a striation light candidate) based on the evaluation function value S obtained by the circuit shown in FIG. This circuit forms a part of the arithmetic circuit 8 shown in FIG. In this figure, 22 and 24 are N comparators, 26 is a register group, and 28 is an ordinal number determination circuit. K is an ordinal number determined by the ordinal number determining circuit 28. Next, the operation of this circuit will be described. As shown in FIG. 1, when the luminance V becomes equal to or less than the threshold value Vo again, the value of the evaluation function value S (= ΣV) is set to N comparators 22.
Is input to In each comparator, a comparison is made with the peak area Si (1 ≦ i ≦ N) obtained earlier in the same line (this Si is a past evaluation function). In the present specification, “past” indicates “past in the same horizontal line data” and does not relate to data before that. Next, “0” in the output (1 bit) from each comparator 22
The number of (S <Si) is counted by the ordinal number determination circuit 28. The output of this circuit 28 is the ordinal number K corresponding to the area of the new peak.
Becomes Further, the ordinal numbers Ki and K obtained in the past are compared by the comparator 24. At this time, if S is smaller than Si (Ki <
K), do not change the state in the register. Similarly, when K = N, the state in the register is not changed. When K ≦ Ki <N−1, Ki is increased by one. Then, when K <N and Ki = N−1 (only one case of 1 ≦ i ≦ N is present), Ki, S
Rewrite i, ST′i, Toi with K, S, ST ′, To, respectively. That is, in the circuit shown in FIG. 4 composed of the respective comparators and registers, if the newly observed striation light candidate is higher (larger in area) than that previously in the register, This is equivalent to lowering your rank by one. Also, if the new streak is lower, the order is not affected, and the data in the register is retained. In addition, new data is overwritten in the register which will be swept out by lowering the rank, and up to N pieces of data obtained from data having a larger area out of the data previously obtained in one horizontal line are always written. Continue to be held. FIG. 5 shows such a sorting process in the form of a program. In the circuit shown in FIG. 4, N independent circuits simultaneously input and perform parallel processing, and required processing can be completed in a few clocks. Note that the logic of the circuit shown here is just an example,
Ki is the peak of the striation light with the largest data of O,
It becomes smaller as the value of Ki becomes larger, and becomes smallest when Ki is N-1. However, this may be changed as necessary, the magnitude of the streak may increase with the value of Ki, or the value of Ki may take a value of 1 to N. In that case, the logic shown in FIGS. 4 and 5 only slightly changes. [Effects of the Invention] The present invention works as a pre-processing circuit of a distance information input device that projects a plurality of line lights on an object, and extracts a necessary line light position and sorts the data in real time, thereby performing a subsequent measurement. The system and processing system can be simplified, the processing time can be significantly reduced, and the storage capacity can be reduced. In addition, not only the use of the distance information input device for preprocessing,
It is also effective for application to extraction of bar code readers and the like.
The signal to be processed need not be a two-dimensional TV image as in this example, but may be a one-dimensional signal.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a luminance level of a streak light displayed on a TV image, FIG. 2 is a block diagram showing an entire embodiment of the present invention, FIG. FIG. 4 shows the operation unit 8 shown in FIG.
FIG. 5 is a block diagram showing the logic of sorting in detail. 2 A / D conversion unit 4 Threshold calculation unit 6 1-line delay unit 8 Operation unit 10 Center-of-gravity calculation unit

Claims (1)

(57) [Claims] Means for determining a protruding signal component appearing on an image signal representing a plurality of linear light as a candidate for linear light, and an evaluation function value for eliminating false linear light candidates for each of the linear light candidates. A means for comparing the evaluation function value previously obtained with the newly obtained evaluation function value; and, as a result of the comparison, only linear light candidates within a predetermined number. A striation light extraction circuit, which processes an image signal at a video rate. 2. 2. The linear light extraction circuit according to claim 1, wherein the area of the protruding signal component is used as the evaluation function value. 3. Means for determining a protruding signal component appearing on an image signal representing a plurality of linear light as a candidate for linear light, and an evaluation function value for eliminating false linear light candidates for each of the linear light candidates. A means for comparing the evaluation function value previously obtained with the newly obtained evaluation function value; and, as a result of the comparison, only linear light candidates within a predetermined number.残 す (V · T) / ΣV, using the luminance information V and the time information T of the pixel points constituting the candidate signal portion from the portion of the image signal that is a candidate for the streak light. A striation light extraction circuit comprising: a weighted average value calculation circuit for determining a striation light position by calculation; and processing an image signal at a video rate.