JPS5920592B2 - Method for detecting defects on the surface of yarn rolls - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for detecting defects on the surface of yarn rolls such as parts, corn, cheese, etc.

When such yarn rolls are shipped as products, appearance inspections are conducted to remove any defects on the surface that make them unsuitable for use as products. Currently, there is no optimal means for automatically and accurately detecting and identifying defects present on the surface of a rolled product that are unsuitable for the product, and thus visual judgment is currently relied on in most cases.

Therefore, the inspection process requires a large number of skilled operators, and since it relies on manual labor, work efficiency is low, and there are individual differences among operators, making quantitative quality control difficult. In view of the current situation, the present inventors focused on a solid-state camera that can photoelectrically scan the surface of a yarn roll at high speed as a method for detecting such defects existing on the surface of a yarn roll, and developed a method for detecting such defects on the surface of a yarn roll. As a result of intensive research on processing methods,
In particular, the present invention is capable of detecting defects on the surface of yarn rolls of complicated shapes such as parts.

That is, when detecting defects on the surface of a yarn roll, the present invention irradiates the surface of the yarn roll to be inspected with light, receives the reflected light with a solid-state camera, and detects the defects every scan of the solid-state image sensor constituting the solid-state camera. Or each element output (V
01 to VOn), and each element output (Vn01 to VnOn) at the nth sampling and each element output at the n+1st sampling (Vn+101 to Vn+
10n) and the output level difference before and after of each element for each sampling vno, -vn+1o7, vno2-vn+, 02,
...VnOn-Vn+, On is determined, and if any difference before and after exceeds the reference value, it is separated as a defect signal, and at each sampling, the difference between the output levels of each element before and after is calculated in the element arrangement direction. Add a certain number at a time and make this 1
A yarn scroll that sequentially calculates values while moving element by element, compares each value with a reference value, separates it as a defect signal when any addition result exceeds the reference signal, and detects a defect based on the presence or absence of the separated defect signal. This is a surface defect detection method.

The present invention as described above makes it possible to automatically and accurately detect various defects present on the surface of a yarn roll. In general, a solid-state camera consists of multiple photo sensors arranged in a linear row or in an area. A TV camera with a built-in solid-state image sensor called a solid-state image sensor.

Each photo sensor, that is, a photoelectric conversion unit, is independent, and the electric signal output corresponding to the amount of incident light of each photoelectric conversion unit (hereinafter referred to as an element) is maintained at a constant level due to the self-scanning function unique to solid-state image sensors. Output to the outside in order. Out of the electric signal outputs (voltage outputs) corresponding to the amount of incident light of each element of the solid-state image sensor, that is, the solid-state camera, which is composed of n elements outputted to the outside in this way, the output of the first successive element is VOl, the nth element output successively output is defined as On. The present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a diagram for explaining an example of the arrangement of the present invention, in which a solid-state camera 3 is installed at a position facing the thread roll 1 to be inspected, with the element arrangement direction being approximately perpendicular to the winding direction of the thread roll 1. The optical system is adjusted and arranged so as to be able to scan the entire width of the thread roll to be inspected, for example, the S section. Also, near the solid-state camera 3, a light source 2 for illuminating the parts,
Place 2'. The light intensity of the light source is set to be sufficiently large within a range that does not saturate the output of the photoelectric conversion section of the solid-state camera. With this arrangement, optical information corresponding to the surface condition of the yarn roll 1 to be inspected is input to each element of the solid-state camera 3, and is sequentially obtained as an electrical signal output. In addition, the yarn roll to be inspected 1
If it is rotated as shown by the arrow, it is possible to obtain the output of the solid-state camera 3 as information converted into an electrical signal corresponding to the state of the entire surface. Solid-state camera 3 used here
Although an area type solid-state image sensor may be used, a linear type is preferable, and a large number of elements is preferable, but it is sufficient if the resolution is at least slightly smaller than the thread diameter. Using a solid-state camera in this way allows scanning to be controlled according to the signal processing speed, and since the elements are independent, the address of the information to be read out is clear, which is advantageous from the standpoint of designing signal processing circuits, etc. It is. A specific method for sampling the output of each element every multiple times of scanning is as follows. That is, generally, a start pulse signal and a clock pulse signal are input as basic signals for driving the solid-state image sensor. Here, the start pulse signal is a signal instructing the start of scanning, and the clock pulse signal is a scanning signal, which supplies a signal with a number of pulses corresponding to the number of elements. The start pulse provides one pulse at the start of each scan. Therefore, count this start pulse,
Information for one scan is sampled by a clock pulse signal using a start pulse signal as a reference signal from a signal obtained by binarizing (digitizing) each element output by an A/D converter every fixed count. Next, Figure 2 shows
This is a block diagram showing an example of a processing circuit that processes electrical signal information corresponding to the surface condition of a yarn roll taken out from a solid-state camera, and a circuit that outputs the results, and the solid-state camera is controlled by a camera control unit 4. 3 analog signal output is buffer amplifier 5
The signal is supplied to the A/D converter via the A/D converter and converted into a digital signal.

The A/D converter 6 used here may be of 6 to 8 bits. Next, the circuit output is sent to a shift register 8 which is formed by serially arranging a number of memories corresponding to the number of elements, and is sent to a shift register 8 which is formed by serially arranging memories corresponding to the number of elements. The information V, Ol~, On (of which the one-dot chain line in the figure is an example) is temporarily stored.
Next, by rotating the thread roll 1, the scanning position of the thread roll surface of each element moves to the second scanning position, and the first element output V2O at this time is sent to the register 8. Therefore, from register 8, the first element information V, O
, is taken out after overflowing and is supplied to the subtraction circuit 10. At this time, the first element output V2O is supplied to the subtraction circuit 10 by opening the gate 9. The subtraction circuit 10 calculates the difference between these two pieces of information, 01−V2Ol,,O2
-V2O2,... -10n-VnOO
is calculated. That is, in this way, the differences between the outputs of all the elements in the time axis direction are sequentially calculated, and the results are sequentially supplied to the discrimination circuit 11. The discrimination circuit 11 compares each piece of information with a reference value set in advance as an allowable limit for defects, and outputs a defect signal only when the allowable limit is exceeded. In this manner, point-type defects or long-line defects in the element arrangement direction can be sharply separated by the front-back difference processing in the time axis direction for each element. On the other hand, the difference between each element in the time axis direction of the subtraction circuit 10 is sequentially supplied to a shift register 12 having four areas arranged in parallel.
2, when the first to fourth pieces of information are first input to all areas, the individual pieces of information are sent to the adder circuit 13, the sum of all the pieces of information is calculated, and the result is sent to the discrimination circuit 14.

Further, only the first information is cleared by the fifth information, and the second to fifth information are added by the addition circuit 13. In this way, the sum of the four pieces of information (hereinafter referred to as moving sum) is calculated each time while sequentially fetching and clearing old information. Next, each moving sum sent to the discrimination circuit 14 is sequentially compared with a reference moving sum set in advance as an allowable limit for defects, and only when the allowable limit is exceeded, a signal is output as a defect. By performing such differential movement processing in the element arrangement direction of the front and back differences, it is possible to sharply separate area-type defects, such as dirt that gradually becomes darker in the inspection width direction, for example. In this way, by using two types of signal processing, it is possible to separate all the defects that appear in the density changes on the surface of the yarn roll.

The controller 7 is a logic circuit for loading data into the registers 8 and 12 and controlling the gate 9 based on the scanning pulse, start pulse, and end pulse from the camera controller 4. Next, the defect signal obtained by calculating the front and rear difference in the discrimination circuit 11 and the defect signal obtained by calculating the moving sum of the front and rear difference are, for example, generated by the alarm mechanism 1
5, and if any processing circuit detects a defect exceeding the allowable limit for the product, it is displayed as a defective product. If the defect signal can be separated in this way, it can be used not only as an alarm but also as a control signal for marking defective yarn rolls or removing defective yarn rolls from the line. By inputting the signals to a computer or the like, it is also possible to aggregate or analyze signals of defects by product type. As described above, the present invention contributes to the automation of the inspection process of yarn rolls and the quantification of quality control, and its effects are extremely large.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram showing an example of the arrangement of the apparatus of the present invention, in which 1 is a yarn roll to be inspected, 2 and 2' are light sources, 3 is a solid-state camera, and S
2 shows an inspection width, and FIG. 2 is a program diagram showing an example of the configuration of the present invention, where 3' is a photoelectric conversion section of a solid-state camera, 4 is a camera controller, 5 is a buffer,
6 is an A/D converter, 7 is a controller, 8 is a shift register, 9 is a gate, 10 is a subtraction circuit, 11 is a discrimination circuit, 12 is a shift register, 13 is an addition circuit, 14 is a discrimination circuit, and 15 is a display circuit. Each is shown below.

Claims (1)

[Claims] 1. When detecting defects on the surface of a yarn roll, the surface of the yarn roll to be inspected is irradiated with light, the reflected light is received by a solid-state camera, and the solid-state image sensor constituting the solid-state camera is scanned every scan or multiple times. Each element output (VO_1 to VOn
), and each element output (VnO_1 to VnOn) at the nth sampling and each element output at the n+1st sampling (V_n_+_1O_1 to V_n_
+_1O_n) and the difference in output level of each element for each sampling V_nO_1-V_n_+_1O_1,V_n
O_2-V_n_+_1O_2,...V_nO_n-V
_n_+_1O_n is determined, and when any difference before and after exceeds the reference value, it is separated as a defect signal, and at each sampling, the difference between the output levels of each element before and after is added by a fixed number in the element arrangement direction, and this is added to one element. The surface of the yarn roll is sequentially obtained while moving one by one, and each value is compared with a reference value. When any of the addition results exceeds the reference signal, it is separated as a defect signal, and a defect is detected by the presence or absence of the separated defect signal. Defect detection method.